Comfy Router’s model catalog - one page of the canonical model IDs that POST /v2/models/{provider}/{model} accepts. An SDK calls this on cold start to discover what is runnable, and the model_not_found suggestions come from the same catalog, so an ID listed here that then 404s on invocation would be worse than either failure alone. That agreement is structural rather than a promise: an entry’s provider and model are the two path segments of the invocation route and reference the same schema components that route’s path parameters do, and id is those two segments joined by /.
Bearer token authentication. Normally a Firebase or Cloud JWT. A 'comfyui-' prefixed API key is also accepted in this header: the prefix classifies the value as an API key and it is validated exactly as if it had been sent in X-API-Key.
The ETag a caller holds from an earlier 200. When it matches the current page (RFC 9110 weak comparison; * matches any current page) the answer is a bodyless 304 carrying the same ETag, otherwise the full page.
Opaque pagination cursor. Pass a previous page's next_cursor to fetch the next page; omit it for the first page. See RouterPageCursor for why the value is opaque and why this route paginates by cursor rather than by offset.
An opaque cursor into a Router list. It is produced by the server and only ever round-tripped: it is not an offset, not a model ID, not ordered, and not stable across catalog rebuilds, so parsing one, incrementing one, or persisting one beyond the walk it came from are all outside the contract. Cursor rather than offset because the catalog is a moving list - an offset walk silently skips or repeats entries when entries are added or removed mid-walk, and a caller cannot tell that it happened.
Number of models to return in one page. Values above the declared maximum are outside the contract, but this route does not reject them: it serves the maximum instead, and the page size actually served is echoed back as limit on the response, so a clamp is always detectable by the caller. Treat the maximum as the real page stride - a client that asks for more and assumes it received more will miss rows. 0 and negative values are also accepted and select the default, which is why no minimum is declared: sub-1 is meaningful here, not invalid.
The page size actually served. A requested limit above the maximum is clamped down to the maximum rather than rejected, so this can be smaller than the value asked for - paginate with this number, not with the one you sent, or you will assume rows you never received.
An opaque cursor into a Router list. It is produced by the server and only ever round-tripped: it is not an offset, not a model ID, not ordered, and not stable across catalog rebuilds, so parsing one, incrementing one, or persisting one beyond the walk it came from are all outside the contract. Cursor rather than offset because the catalog is a moving list - an offset walk silently skips or repeats entries when entries are added or removed mid-walk, and a caller cannot tell that it happened.
{"docsConfig":{"theme":"mint","$schema":"https://mintlify.com/docs.json","name":"ComfyUI","colors":{"primary":"#49378B","light":"#F2FF59","dark":"#F2FF59"},"logo":{"light":"https://mintcdn.com/dripart/WCFdTagRdhtvk9-o/logo/light.svg?fit=max\u0026auto=format\u0026n=WCFdTagRdhtvk9-o\u0026q=85\u0026s=ec112de87239058f5a931a7e4b594965","dark":"https://mintcdn.com/dripart/QzWbjSCBG7w61rR3/logo/dark.svg?fit=max\u0026auto=format\u0026n=QzWbjSCBG7w61rR3\u0026q=85\u0026s=62b4d879fb3aadbd7a468d1539032d6f","href":"/"},"favicon":"https://mintcdn.com/dripart/maKDA164-rrTfyxY/favicon.ico?fit=max\u0026auto=format\u0026n=maKDA164-rrTfyxY\u0026q=85\u0026s=94f3a697cf4c9a1ee3961b69219d0942","appearance":{"default":"dark"},"navbar":{"links":[{"href":"https://blog.comfy.org/?utm_source=docs","label":"Blog"},{"href":"/tutorials/partner-nodes/pricing","label":"Pricing"},{"href":"https://platform.comfy.org/?utm_source=docs","label":"Platform"},{"href":"https://comfy.org/download?utm_source=docs","label":"Download"}],"primary":{"type":"button","label":"Comfy Cloud","href":"https://comfy.org/cloud?utm_source=docs"}},"footer":{"socials":{"github":"https://github.com/Comfy-Org/ComfyUI","x":"https://x.com/ComfyUI","discord":"https://discord.com/invite/comfyorg","youtube":"https://www.youtube.com/@comfyorg"},"links":[{"header":"Resources","items":[{"label":"installation","href":"/installation/system_requirements"},{"label":"Tutorials","href":"/tutorials/basic/text-to-image"},{"label":"Development","href":"/development/overview"}]},{"header":"Products","items":[{"label":"Features","href":"https://www.comfy.org/?utm_source=docs"},{"label":"Gallery","href":"https://www.comfy.org/gallery?utm_source=docs"},{"label":"Download","href":"https://www.comfy.org/download?utm_source=docs"}]},{"header":"Company","items":[{"label":"About","href":"https://www.comfy.org/about?utm_source=docs"},{"label":"Careers","href":"https://www.comfy.org/careers?utm_source=docs"},{"label":"Terms of Service","href":"https://www.comfy.org/terms-of-service?utm_source=docs"},{"label":"Privacy Policy","href":"https://www.comfy.org/privacy-policy?utm_source=docs"}]}]},"search":{"prompt":"Search Comfy documentation..."},"seo":{"metatags":{"googlebot":"max-video-preview:0"}},"icons":{"library":"fontawesome"},"integrations":{"ga4":{"measurementId":"G-MXC01ZDDHQ"},"posthog":{"apiKey":"phc_iKfK86id4xVYws9LybMje0h44eGtfwFgRPIBehmy8rO"}},"contextual":{"options":["copy","view","chatgpt","claude"],"display":"header"},"styling":{"latex":true},"navigation":{"languages":[{"language":"en","href":""},{"language":"zh","footer":{"socials":{"github":"https://github.com/Comfy-Org/ComfyUI","x":"https://x.com/ComfyUI","discord":"https://discord.com/invite/comfyorg","youtube":"https://www.youtube.com/@comfyorg"},"links":[{"header":"资源","items":[{"label":"安装","href":"/zh/installation/system_requirements"},{"label":"教程","href":"/zh/tutorials/basic/text-to-image"},{"label":"开发","href":"/zh/development/overview"}]},{"header":"产品","items":[{"label":"功能","href":"https://www.comfy.org/?utm_source=docs"},{"label":"画廊","href":"https://www.comfy.org/gallery?utm_source=docs"},{"label":"下载","href":"https://www.comfy.org/download?utm_source=docs"}]},{"header":"公司","items":[{"label":"关于","href":"https://www.comfy.org/about?utm_source=docs"},{"label":"招聘","href":"https://www.comfy.org/careers?utm_source=docs"},{"label":"服务条款","href":"https://www.comfy.org/terms-of-service?utm_source=docs"},{"label":"隐私政策","href":"https://www.comfy.org/privacy-policy?utm_source=docs"}]}]},"href":"zh"},{"language":"ja","footer":{"socials":{"github":"https://github.com/Comfy-Org/ComfyUI","x":"https://x.com/ComfyUI","discord":"https://discord.com/invite/comfyorg","youtube":"https://www.youtube.com/@comfyorg"},"links":[{"header":"リソース","items":[{"label":"インストール","href":"/ja/installation/system_requirements"},{"label":"チュートリアル","href":"/ja/tutorials/basic/text-to-image"},{"label":"開発","href":"/ja/development/overview"}]},{"header":"プロダクト","items":[{"label":"機能","href":"https://www.comfy.org/?utm_source=docs"},{"label":"ギャラリー","href":"https://www.comfy.org/gallery?utm_source=docs"},{"label":"ダウンロード","href":"https://www.comfy.org/download?utm_source=docs"}]},{"header":"会社情報","items":[{"label":"概要","href":"https://www.comfy.org/about?utm_source=docs"},{"label":"採用情報","href":"https://www.comfy.org/careers?utm_source=docs"},{"label":"利用規約","href":"https://www.comfy.org/terms-of-service?utm_source=docs"},{"label":"プライバシーポリシー","href":"https://www.comfy.org/privacy-policy?utm_source=docs"}]}]},"navbar":{"links":[{"href":"/ja/tutorials/partner-nodes/pricing","label":"料金"},{"href":"https://comfy.org/download?utm_source=docs","label":"ダウンロード"}],"primary":{"type":"button","label":"Comfy Cloud","href":"https://comfy.org/cloud?utm_source=docs"}},"href":"ja"},{"language":"ko","footer":{"socials":{"github":"https://github.com/Comfy-Org/ComfyUI","x":"https://x.com/ComfyUI","discord":"https://discord.com/invite/comfyorg","youtube":"https://www.youtube.com/@comfyorg"},"links":[{"header":"리소스","items":[{"label":"설치","href":"/ko/installation/system_requirements"},{"label":"튜토리얼","href":"/ko/tutorials/basic/text-to-image"},{"label":"개발","href":"/ko/development/overview"}]},{"header":"제품","items":[{"label":"기능","href":"https://www.comfy.org/?utm_source=docs"},{"label":"갤러리","href":"https://www.comfy.org/gallery?utm_source=docs"},{"label":"다운로드","href":"https://www.comfy.org/download?utm_source=docs"}]},{"header":"회사 정보","items":[{"label":"소개","href":"https://www.comfy.org/about?utm_source=docs"},{"label":"채용","href":"https://www.comfy.org/careers?utm_source=docs"},{"label":"이용약관","href":"https://www.comfy.org/terms-of-service?utm_source=docs"},{"label":"개인정보 처리방침","href":"https://www.comfy.org/privacy-policy?utm_source=docs"}]}]},"navbar":{"links":[{"href":"/ko/tutorials/partner-nodes/pricing","label":"가격 정책"},{"href":"https://comfy.org/download?utm_source=docs","label":"다운로드"}],"primary":{"type":"button","label":"Comfy Cloud","href":"https://comfy.org/cloud?utm_source=docs"}},"href":"ko"},{"language":"es","href":"es"}]}},"deployment":{"deploymentMetadataValue":{"subdomain":"dripart","hasAudioNarration":false,"actualSubdomain":"dripart","customDomains":["docs.comfy.org"],"gitSource":{"type":"github","owner":"comfy-org","repo":"docs","deployBranch":"main","contentDirectory":"","isPrivate":false,"mountPath":"","label":""},"feedback":{"thumbs":true,"edits":true,"issues":true,"agentFeedback":true},"entitlements":{"AI_CHAT":{"status":"DISABLED"},"WEB_MCP":{"status":"ENABLED"},"NARRATION":{"status":"DISABLED"}},"buildId":"6af234bf4fe798504546db36","clientVersion":"0.0.4094","searchFilterCounts":[{"language":"en","tag":"Built-in Nodes","count":1166},{"language":"ja","tag":"組み込みノード","count":1166},{"language":"zh","tag":"内置节点","count":1166},{"language":"en","tag":"Nodes","count":1165},{"language":"ja","tag":"ノード","count":1165},{"language":"zh","tag":"节点","count":1165},{"language":"ko","tag":"내장 노드 (Built-in Nodes)","count":1124},{"language":"ko","tag":"노드","count":1123},{"language":"es","tag":"Nodos integrados","count":1058},{"language":"es","tag":"Nodos","count":1057},{"language":"en","tag":"Tutorials","count":419},{"language":"es","tag":"Tutoriales","count":419},{"language":"en","tag":"Model","count":381},{"language":"ja","tag":"Model","count":381},{"language":"ko","tag":"Model","count":381},{"language":"zh","tag":"Model","count":381},{"language":"es","tag":"Modelo","count":378},{"language":"en","tag":"Image","count":355},{"language":"en","tag":"Partner","count":348},{"language":"ja","tag":"Partner","count":348},{"language":"ko","tag":"Partner","count":348},{"language":"zh","tag":"Partner","count":348},{"language":"es","tag":"Partner","count":346},{"language":"ja","tag":"API 開発","count":326},{"language":"ko","tag":"API 개발","count":326},{"language":"zh","tag":"API 开发","count":326},{"language":"es","tag":"Imagen","count":325},{"language":"ja","tag":"はじめに","count":312},{"language":"ko","tag":"시작하기","count":312},{"language":"zh","tag":"开始使用","count":312},{"language":"en","tag":"Video","count":291},{"language":"en","tag":"Comfy Router","count":289},{"language":"ja","tag":"Comfy Router","count":289},{"language":"ko","tag":"Comfy Router","count":289},{"language":"zh","tag":"Comfy Router","count":289},{"language":"es","tag":"Video","count":266},{"language":"es","tag":"Comfy Router","count":266},{"language":"en","tag":"Models","count":259},{"language":"ja","tag":"Models","count":259},{"language":"ko","tag":"Models","count":259},{"language":"zh","tag":"Models","count":259},{"language":"en","tag":"All Models","count":256},{"language":"es","tag":"Models","count":227},{"language":"es","tag":"All Models","count":224},{"language":"ja","tag":"画像","count":207},{"language":"ko","tag":"이미지","count":207},{"language":"zh","tag":"图像","count":207},{"language":"ja","tag":"Video","count":193},{"language":"zh","tag":"Video","count":193},{"language":"ko","tag":"Video","count":176}],"searchSettings":{"maxResults":10}},"assistantConfig":{"enableStarterQuestions":false},"basePath":"","hasChatPermissions":false,"hasWebMcpPermissions":true,"toggles":[{"name":"kill-isr","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"agent-mintlify-slack","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"agent-integrations-settings","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"rbac-expansion","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"automations-pr-grouping","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"custom-prefix","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"analytics-v7","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"user-flows-analytics","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"no-publish-flow","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-onboarding","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"llms-txt-v2","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"snapshot-serving-live","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"stripTemperatureMiddleware","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"support-platform-pylon-zendesk","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"snapshot-serving","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"instant-rollback","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"granular-notifications-server","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"granular-notifications","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"automations-fable","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"fixed-ai-action-pricing","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"automation-resolve-merge-conflict","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"scoped-publishing","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"mintql","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"page-level-analytics","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"editor-version-history-v2","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"billing-page-v2","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"full-page-search","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workspace-projects","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"assistant-v5","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"fork-pr-previews","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"revalidation-jwt","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-sso-scim","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"hobby-abuse-scan","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"dashboard-inbox","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-signup","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"turbopuffer-dual-write","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-enabled","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"approve-review-flow","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"editor-metadata","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"sso-setup-deadline-banner","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"turbopuffer-serve","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"live-agents","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"comment-mode","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false}],"googleFontsInlined":false},"customCss":[{"filePath":"custom.css","content":":root{--ink-900: #151417;--hub-shell-light: #f7f6f9;--accent-purple: #49378b;--accent-yellow: #f2ff59;--primary: 73 55 139 !important;--primary-light: 242 255 89 !important;--primary-dark: 242 255 89 !important}html.dark{--primary: 242 255 89 !important;--background-dark: 14 14 13 !important}html.dark body{background-color:#0e0e0d!important}html.dark #navbar,html.dark #navbar-transition{background-color:var(--ink-900)!important}html.dark #sidebar,html.dark #sidebar-content{background-color:rgb(var(--background-dark))!important}html.dark #navbar,html.dark #navbar-transition{box-shadow:none!important;filter:none!important;backdrop-filter:none!important;-webkit-backdrop-filter:none!important}html.dark #navbar-transition{border-bottom:1px solid rgba(255, 255, 255, 0.08)!important;opacity:1!important}html.dark img.nav-logo{filter:none}html.dark #sidebar-content a:not([aria-current=\"page\"]){color:rgba(255, 255, 255, 0.72)!important}html.dark #sidebar-content a:not([aria-current=\"page\"]) *{color:inherit!important}html.dark #sidebar-content a[aria-current=\"page\"]{color:var(--accent-yellow)!important;background-color:rgba(242, 255, 89, 0.10)!important}html.dark #sidebar-content a[aria-current=\"page\"] *{color:inherit!important}html.dark .nav-tabs-item[data-active],html.dark mobile-nav-tabs-item[data-active]{color:var(--accent-yellow)!important}html.dark #navbar a[href*=\"comfy.org/cloud\"] span.absolute{background-color:var(--accent-yellow)!important}html.dark #navbar a[href*=\"comfy.org/cloud\"] span.text-white{color:#151417!important}html:not(.dark) #background-color,html:not(.dark) #navbar,html:not(.dark) #navbar-transition,html:not(.dark) #sidebar,html:not(.dark) #sidebar-content{background-color:var(--hub-shell-light)!important}html:not(.dark) .nav-tabs-item[data-active],html:not(.dark) mobile-nav-tabs-item[data-active],html:not(.dark) nav-dropdown-item[data-active]{background-color:transparent!important;box-shadow:none!important;color:var(--accent-purple)!important;font-weight:600!important}html:not(.dark) .nav-tabs-item[data-active] *,html:not(.dark) mobile-nav-tabs-item[data-active] *,html:not(.dark) nav-dropdown-item[data-active] *{color:inherit!important}html:not(.dark) #sidebar-content li:has(>a[aria-current=\"page\"]){background-color:transparent!important;box-shadow:none!important}html:not(.dark) #sidebar-content a[aria-current=\"page\"]{color:var(--accent-purple)!important;background-color:rgba(73, 55, 139, 0.1)!important;font-weight:600!important}html:not(.dark) #sidebar-content a[aria-current=\"page\"] *{color:inherit!important}html:not(.dark) toc-item[data-active],html:not(.dark) toc-item[data-active-deepest],html:not(.dark) toc-item[data-active] a,html:not(.dark) toc-item[data-active-deepest] a{background-color:transparent!important;color:var(--accent-purple)!important;font-weight:600!important}html:not(.dark) #header :where(h1,h2,h3,h4,h5,h6),html:not(.dark) #content-area :where(h1,h2,h3,h4,h5,h6){color:#151417!important}html:not(.dark) #navbar a[href*=\"comfy.org/cloud\"]{background-color:transparent!important;color:var(--accent-purple)!important;border-radius:0.75rem;box-shadow:inset 0 0 0 1px var(--accent-purple)!important}html:not(.dark) #navbar a[href*=\"comfy.org/cloud\"] span.absolute{background-color:transparent!important}html:not(.dark) #navbar a[href*=\"comfy.org/cloud\"] *{color:inherit!important}#sidebar .sticky.top-0.pointer-events-none.bg-linear-to-b{background-image:none!important}@media (min-width: 1024px){#navbar.is-not-custom.is-not-center+div:has(>#sidebar),.max-w-8xl.mx-auto.relative,.max-w-8xl.mx-auto{max-width:none!important;width:100%!important;margin-left:0!important;margin-right:0!important;padding-left:0!important;padding-right:0!important}#navbar [class*=\"px-12\"],#navbar [class*=\"mx-4\"]{padding-left:0!important;padding-right:1.5rem!important;margin-left:0!important;margin-right:0!important}#navbar a.select-none,#navbar .nav-logo{margin-left:1rem!important}#navbar .nav-tabs{padding-left:2rem!important}#sidebar{left:0!important;margin-left:0!important;box-shadow:none!important}#sidebar-content{margin-left:0!important;padding-left:1rem!important;box-sizing:border-box!important}#navbar,#navbar-transition{left:0!important;right:0!important;width:100%!important;max-width:none!important}#content-area{margin-left:0!important;max-width:none!important;width:100%!important;padding-left:18rem!important}#sidebar.sticky~#content-container #content-area{padding-left:2rem!important;padding-right:2rem!important}#content-container,#body-content{max-width:none!important;width:100%!important}:root{--content-width: 100%;--sidebar-width: 18rem}}@media (min-width: 1024px){nav#sidebar,#sidebar[style]{left:0!important;right:auto!important;margin-left:0!important;transform:none!important}#navbar li.navbar-link:has(a[href*=\"comfy.org/download\"]){margin-right:0!important}#navbar a[href*=\"comfy.org/download\"]{box-sizing:border-box;border:1px solid rgb(255 255 255 / 0.22);border-radius:0.75rem;padding:7px 0.875rem!important;color:#fff!important;font-size:0.875rem;line-height:1.25rem}html:not(.dark) #navbar a[href*=\"comfy.org/download\"]{border-color:rgb(0 0 0 / 0.16);color:#151417!important}#navbar a[href*=\"comfy.org/download\"]:hover{background:rgb(255 255 255 / 0.06)}html:not(.dark) #navbar a[href*=\"comfy.org/download\"]:hover{background:rgb(0 0 0 / 0.04)}#navbar li.navbar-link:has(a[href*=\"comfy.org/download\"])+.whitespace-nowrap,#navbar .flex.space-x-6>.whitespace-nowrap.hidden{margin-left:0.5rem!important}}body:has(.router-api-reference-marker) #content,body:has(.router-api-reference-marker) #header{max-width:48rem}body:has(.router-api-reference-marker) #content-area table{width:100%;table-layout:auto}body:has(.router-api-reference-marker) #content-area th,body:has(.router-api-reference-marker) #content-area td{min-width:4rem!important;overflow-wrap:anywhere}body:has(.router-api-reference-marker) #content-area table:has(th:nth-child(6)) th:nth-child(3),body:has(.router-api-reference-marker) #content-area table:has(th:nth-child(6)) td:nth-child(3){min-width:5rem!important}@media (max-width: 1535px){body:has(.router-api-reference-marker) #content-side-layout{display:none!important}}body:has(.router-provider-coverage-marker) #content,body:has(.router-provider-coverage-marker) #header{max-width:64rem}body:has(.router-provider-coverage-marker) #content-area table{width:100%;table-layout:auto}body:has(.router-provider-coverage-marker) #content-area th,body:has(.router-provider-coverage-marker) #content-area td{min-width:5rem!important;overflow-wrap:anywhere}body:has(.router-provider-coverage-marker) #content-area th:first-child,body:has(.router-provider-coverage-marker) #content-area td:first-child{min-width:12rem!important}body:has(.router-quickstart-marker) [data-component-part=\"card-content\"] [data-component-part=\"code-block-root\"][style*=\"height:190px\"],body:has(.router-quickstart-marker) [data-component-part=\"card-content\"] [data-component-part=\"code-block-root\"][style*=\"height: 190px\"]{height:86px!important}"},{"filePath":"samples-cards.css","content":".sample-card-grid{display:grid;gap:20px;grid-template-columns:repeat(2, minmax(0, 1fr));margin-top:28px}.sample-card{min-width:0;perspective:1200px}.sample-card-inner{height:565px;position:relative;transform-style:preserve-3d;transition:transform 560ms cubic-bezier(.2, .75, .3, 1)}.sample-card[data-flipped=\"true\"] .sample-card-inner{transform:rotateY(180deg)}.sample-card-face{backface-visibility:hidden;background:#fff;border:1px solid #e5e1eb;border-radius:12px;box-sizing:border-box;height:100%;inset:0;overflow:hidden;position:absolute;width:100%}.sample-card-front{display:flex;flex-direction:column}.sample-card-back{opacity:0;pointer-events:none;transform:rotateY(180deg);visibility:hidden}.sample-card[data-flipped=\"true\"] .sample-card-front{pointer-events:none;transition:visibility 0s linear 280ms;visibility:hidden}.sample-card[data-flipped=\"true\"][data-active-path=\"api\"] .sample-card-back[data-sample-path=\"api\"],.sample-card[data-flipped=\"true\"][data-active-path=\"endpoint\"] .sample-card-back[data-sample-path=\"endpoint\"]{opacity:1;pointer-events:auto;visibility:visible}.sample-card-preview{border-bottom:1px solid #e5e1eb;cursor:pointer;display:block;width:100%}.sample-card-body{display:flex;flex:1;flex-direction:column;padding:20px}.sample-card-body h2,.sample-path-body h2{color:#49378b;font-size:18px;font-weight:700;letter-spacing:-.02em;margin:0 0 8px}.sample-path-body h2{text-align:center}.sample-card-body>p{color:#5f5966;font-size:14px;line-height:1.55;margin:0;min-height:66px}.sample-card-actions{display:flex;gap:8px;margin-top:20px}.sample-card-actions button{background:#fff;border:1px solid #cfc8de;border-radius:6px;color:#49378b;cursor:pointer;flex:1;font-size:13px;font-weight:700;padding:10px 12px}.sample-card-links{border-top:1px solid #eeeaf1;display:flex;gap:8px;margin-top:16px;padding-top:12px}.sample-card-links a{color:#625b6c;font-size:12px;font-weight:650;text-decoration:none}.sample-card-links span{color:#b3adba}.sample-path-body{padding:20px}.sample-path-body>button{background:transparent;border:0;color:#625b6c;cursor:pointer;font-size:12px;font-weight:700;padding:0}.sample-path-eyebrow{color:#857a9e;font-size:10px;font-weight:700;letter-spacing:.08em;margin:19px 0 5px;text-align:center;text-transform:uppercase}.sample-path-body>p:not(.sample-path-eyebrow){color:#5f5966;font-size:12px;line-height:1.5;margin:0 0 15px}.sample-path-steps{display:grid;gap:8px}.sample-path-step{align-items:center;background:#faf9fc;border:1px solid #e5e1ea;border-radius:8px;display:flex;gap:10px;padding:10px 11px}.sample-path-number{align-items:center;background:#49378b;border-radius:999px;color:#fff;display:inline-flex;flex:0 0 21px;font-size:11px;font-weight:700;height:21px;justify-content:center}.sample-path-copy{flex:1;min-width:0}.sample-path-copy strong{color:#312c36;display:block;font-size:12px}.sample-path-copy span{color:#706a75;display:block;font-size:11px;line-height:1.35;margin-top:2px}.sample-path-step a{color:#49378b;flex:0 0 auto;font-size:11px;font-weight:700;text-decoration:none}@media (max-width: 768px){.sample-card-grid{grid-template-columns:1fr}.sample-card-inner{height:clamp(620px, 180vw, 760px)}}"}],"page":{"pagePath":"zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run","slug":"/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run","pageMetadata":{"title":"List the models Comfy Router can run.","description":"Comfy Router's model catalog - one page of the canonical model IDs that `POST /v2/models/{provider}/{model}` accepts. An SDK calls this on cold start to discover what is runnable, and the `model_not_found` suggestions come from the same catalog, so an ID listed here that then 404s on invocation would be worse than either failure alone. That agreement is structural rather than a promise: an entry's `provider` and `model` are the two path segments of the invocation route and reference the same schema components that route's path parameters do, and `id` is those two segments joined by `/`.","deprecated":null,"version":null,"openapi":"/router-openapi.yaml get /v2/models","href":"/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run","autogeneratedByOpenApi":true},"description":{"compiledSource":"\"use strict\";\nconst {jsx: _jsx, jsxs: _jsxs} = arguments[0];\nconst {useMDXComponents: _provideComponents} = arguments[0];\nfunction _createMdxContent(props) {\n const _components = {\n code: \"code\",\n p: \"p\",\n ..._provideComponents(),\n ...props.components\n };\n return _jsxs(_components.p, {\n children: [\"Comfy Router’s model catalog - one page of the canonical model IDs that \", _jsx(_components.code, {\n children: \"POST /v2/models/{provider}/{model}\"\n }), \" accepts. An SDK calls this on cold start to discover what is runnable, and the \", _jsx(_components.code, {\n children: \"model_not_found\"\n }), \" suggestions come from the same catalog, so an ID listed here that then 404s on invocation would be worse than either failure alone. That agreement is structural rather than a promise: an entry’s \", _jsx(_components.code, {\n children: \"provider\"\n }), \" and \", _jsx(_components.code, {\n children: \"model\"\n }), \" are the two path segments of the invocation route and reference the same schema components that route’s path parameters do, and \", _jsx(_components.code, {\n children: \"id\"\n }), \" is those two segments joined by \", _jsx(_components.code, {\n children: \"/\"\n }), \".\"]\n });\n}\nfunction MDXContent(props = {}) {\n const {wrapper: MDXLayout} = {\n ..._provideComponents(),\n ...props.components\n };\n return MDXLayout ? _jsx(MDXLayout, {\n ...props,\n children: _jsx(_createMdxContent, {\n ...props\n })\n }) : _createMdxContent(props);\n}\nreturn {\n default: MDXContent\n};\n","frontmatter":{},"scope":{}},"pageType":"default","mdx":{"mdxSource":{"compiledSource":"\"use strict\";const{Fragment:_Fragment,jsx:_jsx}=arguments[0];const{useMDXComponents:_provideComponents}=arguments[0];function _createMdxContent(props){return _jsx(_Fragment,{})}function MDXContent(props={}){const{wrapper:MDXLayout}={..._provideComponents(),...props.components};return MDXLayout?_jsx(MDXLayout,{...props,children:_jsx(_createMdxContent,{...props})}):_createMdxContent(props)}return{default:MDXContent}","frontmatter":{},"scope":{"config":{},"pageMetadata":{"title":"List the models Comfy Router can run.","description":"Comfy Router's model catalog - one page of the canonical model IDs that `POST /v2/models/{provider}/{model}` accepts. An SDK calls this on cold start to discover what is runnable, and the `model_not_found` suggestions come from the same catalog, so an ID listed here that then 404s on invocation would be worse than either failure alone. That agreement is structural rather than a promise: an entry's `provider` and `model` are the two path segments of the invocation route and reference the same schema components that route's path parameters do, and `id` is those two segments joined by `/`.","deprecated":null,"version":null,"openapi":"/router-openapi.yaml get /v2/models","href":"/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run","autogeneratedByOpenApi":true}}},"mdxSourceWithNoJs":null},"mdxExtracts":{"tableOfContents":[],"codeExamples":{}},"jsLinks":["/_mintlify/custom/dripart/9d6bf2b2531e904c/giscus-comments.js","/_mintlify/custom/dripart/bd0341946f3c75a6/syft.js","/_mintlify/custom/dripart/042392a43b5e681e/samples-cards.js"],"openApiReferenceData":{"operation":{"summary":"List the models Comfy Router can run.","description":"Comfy Router's model catalog - one page of the canonical model IDs that `POST /v2/models/{provider}/{model}` accepts. An SDK calls this on cold start to discover what is runnable, and the `model_not_found` suggestions come from the same catalog, so an ID listed here that then 404s on invocation would be worse than either failure alone. That agreement is structural rather than a promise: an entry's `provider` and `model` are the two path segments of the invocation route and reference the same schema components that route's path parameters do, and `id` is those two segments joined by `/`.","operationId":"listRouterModels","tags":["Comfy Router"],"security":["e7e3f8e5-bea5-4030-b1db-8f46dca76241","3ad52369-c0be-4ebc-8842-96afd522dd68"],"parameters":["346128f7-1e08-4d17-8c05-ecc3f425fc66","23a7330f-1884-4d27-8f45-d604b1d5c617","82476bc5-2879-454f-9d6e-1845d50da289"],"responses":{"200":"8120725a-2821-40b4-bfdb-3ee7d81a7305","304":"2543534e-876a-4932-9cef-672c141419f2","400":"2fa33a34-4b23-4d67-8f34-fe19c54fc1d7","401":"91a10445-9225-4d4d-a974-7c7bf264501a","403":"3171abdf-95c0-44af-b054-6eac9e52f872","429":"7dfa64fa-4be2-4272-8d6a-a542b27c997b","503":"a2bbb60a-4833-42cc-a11e-be4f6a942588"},"method":"get","type":"path","servers":["bf9722b0-5952-4eeb-9e69-c43ede83ce5a"],"path":"/v2/models","baseUrlOptions":["https://api.comfy.org"],"title":"List the models Comfy Router can run.","codeSamples":[],"requiredOnlyExamples":false,"requestExampleType":"generated","responseExampleType":"generated"},"dependencies":{"parameters":{"346128f7-1e08-4d17-8c05-ecc3f425fc66":{"name":"cursor","in":"query","required":false,"description":"Opaque pagination cursor. Pass a previous page's `next_cursor` to fetch the next page; omit it for the first page. See `RouterPageCursor` for why the value is opaque and why this route paginates by cursor rather than by offset.","schema":{"type":"string","description":"Opaque pagination cursor. Pass a previous page's `next_cursor` to fetch the next page; omit it for the first page. See `RouterPageCursor` for why the value is opaque and why this route paginates by cursor rather than by offset.\nAn opaque cursor into a Router list. It is produced by the server and only ever round-tripped: it is not an offset, not a model ID, not ordered, and not stable across catalog rebuilds, so parsing one, incrementing one, or persisting one beyond the walk it came from are all outside the contract. Cursor rather than offset because the catalog is a moving list - an offset walk silently skips or repeats entries when entries are added or removed mid-walk, and a caller cannot tell that it happened.","pattern":"^[A-Za-z0-9._~+/=-]+$","minLength":1,"maxLength":512,"example":"q7Fm2xTn9pLd4RsV","isRequired":false,"uniqueKey":"","typeLabel":"string"}},"23a7330f-1884-4d27-8f45-d604b1d5c617":{"name":"limit","in":"query","required":false,"description":"Number of models to return in one page. Values above the declared maximum are outside the contract, but this route does not reject them: it serves the maximum instead, and the page size actually served is echoed back as `limit` on the response, so a clamp is always detectable by the caller. Treat the maximum as the real page stride - a client that asks for more and assumes it received more will miss rows. 0 and negative values are also accepted and select the default, which is why no `minimum` is declared: sub-1 is meaningful here, not invalid.","schema":{"type":"integer","maximum":100,"default":20,"isRequired":false,"description":"Number of models to return in one page. Values above the declared maximum are outside the contract, but this route does not reject them: it serves the maximum instead, and the page size actually served is echoed back as `limit` on the response, so a clamp is always detectable by the caller. Treat the maximum as the real page stride - a client that asks for more and assumes it received more will miss rows. 0 and negative values are also accepted and select the default, which is why no `minimum` is declared: sub-1 is meaningful here, not invalid.","uniqueKey":"","typeLabel":"integer"}},"82476bc5-2879-454f-9d6e-1845d50da289":{"in":"header","name":"If-None-Match","required":false,"description":"The `ETag` a caller holds from an earlier `200`. When it matches the current page (RFC 9110 weak comparison; `*` matches any current page) the answer is a bodyless `304` carrying the same `ETag`, otherwise the full page.","schema":{"type":"string","isRequired":false,"description":"The `ETag` a caller holds from an earlier `200`. When it matches the current page (RFC 9110 weak comparison; `*` matches any current page) the answer is a bodyless `304` carrying the same `ETag`, otherwise the full page.","uniqueKey":"","typeLabel":"string"}}},"responses":{"8120725a-2821-40b4-bfdb-3ee7d81a7305":{"description":"OK - one page of the model catalog.","headers":{"X-Comfy-Request-Id":{"description":"Server-generated identifier for this call, present on every Router response - success, 4xx and 5xx alike, because an error response is exactly when a user needs an id to quote in a support request. The same value is written into the call's usage/audit event, which is what lets a complaint about a charge be joined to the charge itself instead of searched for by timestamp.","required":true,"schema":{"type":"string","format":"uuid","example":"6f1a1a6e-6a53-4a5f-9d3a-2b3b0a1f9c21"}},"ETag":{"description":"Strong entity tag over the served body's bytes, for `GET /v2/models` and `GET /v2/models/{provider}/{model}`. A caller that re-reads the catalog should store this value and send it back as `If-None-Match` to get a `304` instead of the body.","required":true,"schema":{"type":"string","example":"\"6b8c1f2e0a9d4c3b5e7f8a1b2c3d4e5f60718293a4b5c6d7e8f90a1b2c3d4e5f\""}},"Cache-Control":{"description":"Freshness directives for a catalog page or entry. `private` because the answer is decided per credential (the response also carries `Vary: Authorization, X-API-Key`), so a shared cache must not hold it, and `must-revalidate` so a stale copy is revalidated against the `ETag` rather than served on. Error responses on these routes are `private, no-store` instead.","required":false,"schema":{"type":"string","example":"private, max-age=300, must-revalidate"}}},"content":{"application/json":{"schema":{"type":"object","description":"One page of the Router model catalog.","properties":{"data":{"type":"array","description":"The models on this page, at most `limit` of them.","items":{"type":"object","description":"One entry in the Router model catalog: the identity of a runnable model, and nothing else. The per-model detail route composes this same entry rather than restating it, which is why the name is `...ListEntry` and not `...Summary` - there must be exactly one definition of what a catalog entry is. Per-model detail and the per-model input/output schemas are their own routes, so this shape stays the minimum a caller needs in order to invoke the model - deliberately, because this is the payload an SDK fetches on cold start. `id` is `provider` and `model` joined by `/`; the two fields are carried separately as well so a caller composes the invocation path without splitting a string.","properties":{"id":{"type":"string","description":"A canonical Comfy Router model ID, `{provider}/{model}` - exactly the value that addresses the model on `POST /v2/models/{provider}/{model}`, so a caller can interpolate it into that path without re-deriving it from anything. Its `pattern` is `RouterProviderSegment` and `RouterModelSegment` joined by a single `/`, and `maxLength` is their sum plus that separator.","pattern":"^[a-z0-9]+([._-][a-z0-9]+)*/[a-z0-9]+([._-][a-z0-9]+)*$","maxLength":193,"example":"bfl/flux-2-pro","isRequired":true,"uniqueKey":"data.items.[INDEX].id","typeLabel":"string"},"provider":{"type":"string","description":"Lowercase `provider` segment of the canonical `{provider}/{model}` model ID - the partner whose model is being addressed. The invocation route's `provider` path parameter and a catalog entry's `provider` field both reference this one schema, which is what keeps the listed IDs and the accepted IDs from drifting apart.","pattern":"^[a-z0-9]+([._-][a-z0-9]+)*$","maxLength":64,"example":"bfl","isRequired":true,"uniqueKey":"data.items.[INDEX].provider","typeLabel":"string"},"model":{"type":"string","description":"Lowercase `model` segment of the canonical `{provider}/{model}` model ID - the model to run within that provider. Shared by the invocation route's `model` path parameter and a catalog entry's `model` field, for the same no-drift reason as `RouterProviderSegment`.","pattern":"^[a-z0-9]+([._-][a-z0-9]+)*$","maxLength":128,"example":"flux-2-pro","isRequired":true,"uniqueKey":"data.items.[INDEX].model","typeLabel":"string"},"billing":{"type":"object","description":"Per-model billing facts a caller needs before invoking - not prices. Usage and cost figures never appear here.","properties":{"charges_on_policy_rejection":{"type":"string","description":"Whether a call this model refuses on content-policy grounds is nevertheless charged to the caller. Providers differ, the difference is invisible at call time, and a user who sees an error and a charge for the same call has no way to have known - so it is stated per model, before the call, rather than left to per-provider folklore. When it is `yes`, a `content_policy_violation` error on the synchronous route can carry `X-Comfy-Credits-Used`: it does when the provider refused the model's generated output and Comfy recorded usage for that refused generation, and carries that amount. An input-side refusal, a same-key replay (`Idempotent-Replayed: true`), and a charge too small to show in the header's two decimals do not carry it; a refusal without the header is not reported, not proof it was free.","example":"unknown","isRequired":true,"uniqueKey":"data.items.[INDEX].billing.charges_on_policy_rejection","typeLabel":"string"}},"required":["charges_on_policy_rejection"],"isRequired":true,"uniqueKey":"data.items.[INDEX].billing","typeLabel":"object"},"schema_etag":{"type":"string","description":"The current strong `ETag` of this model's `GET /v2/models/{provider}/{model}/openapi.json` document. Send it as `If-None-Match` on that read to get a `304` while the schema is unchanged, or compare it with the value seen last time to learn which models' schemas changed without fetching every document. Absent whenever that read would not answer `200` for this caller at this moment: the schema service is not available, the model's document cannot be rendered, or the model stopped being available to this caller while the response was being built. An absent tag is not a schema change; re-read the model's document to learn its current tag.","example":"\"9f2c1e0b7a4d3c6e8f5a2b1d0c9e8f7a6b5c4d3e2f1a0b9c8d7e6f5a4b3c2d1e\"","uniqueKey":"data.items.[INDEX].schema_etag","typeLabel":"string"}},"required":["id","provider","model","billing"],"uniqueKey":"data.items.[INDEX]","typeLabel":"object"},"isRequired":true,"uniqueKey":"data","typeLabel":"object[]"},"has_more":{"type":"boolean","description":"Whether another page exists beyond this one. Keep walking while this is true; do not infer the end of the catalog from a short or empty `data`.","isRequired":true,"uniqueKey":"has_more","typeLabel":"boolean"},"limit":{"type":"integer","description":"The page size actually served. A requested `limit` above the maximum is clamped down to the maximum rather than rejected, so this can be smaller than the value asked for - paginate with this number, not with the one you sent, or you will assume rows you never received.","minimum":1,"maximum":100,"example":20,"isRequired":true,"uniqueKey":"limit","typeLabel":"integer"},"next_cursor":{"type":"string","description":"An opaque cursor into a Router list. It is produced by the server and only ever round-tripped: it is not an offset, not a model ID, not ordered, and not stable across catalog rebuilds, so parsing one, incrementing one, or persisting one beyond the walk it came from are all outside the contract. Cursor rather than offset because the catalog is a moving list - an offset walk silently skips or repeats entries when entries are added or removed mid-walk, and a caller cannot tell that it happened.","pattern":"^[A-Za-z0-9._~+/=-]+$","minLength":1,"maxLength":512,"example":"q7Fm2xTn9pLd4RsV","uniqueKey":"next_cursor","typeLabel":"string"}},"required":["data","has_more","limit"],"uniqueKey":"","typeLabel":"object"},"examples":{"Example":{"title":"Example","value":{"data":[{"id":"bfl/flux-2-pro","provider":"bfl","model":"flux-2-pro","billing":{"charges_on_policy_rejection":"unknown"},"schema_etag":"\"9f2c1e0b7a4d3c6e8f5a2b1d0c9e8f7a6b5c4d3e2f1a0b9c8d7e6f5a4b3c2d1e\""}],"has_more":true,"limit":20,"next_cursor":"q7Fm2xTn9pLd4RsV"}}},"exampleType":"generated"}},"exampleType":"generated"},"2543534e-876a-4932-9cef-672c141419f2":{"description":"Not Modified - the page is unchanged since the `ETag` the caller sent in `If-None-Match`. No body is returned.","headers":{"X-Comfy-Request-Id":{"description":"Server-generated identifier for this call, present on every Router response - success, 4xx and 5xx alike, because an error response is exactly when a user needs an id to quote in a support request. The same value is written into the call's usage/audit event, which is what lets a complaint about a charge be joined to the charge itself instead of searched for by timestamp.","required":true,"schema":{"type":"string","format":"uuid","example":"6f1a1a6e-6a53-4a5f-9d3a-2b3b0a1f9c21"}},"ETag":{"description":"Strong entity tag over the served body's bytes, for `GET /v2/models` and `GET /v2/models/{provider}/{model}`. A caller that re-reads the catalog should store this value and send it back as `If-None-Match` to get a `304` instead of the body.","required":true,"schema":{"type":"string","example":"\"6b8c1f2e0a9d4c3b5e7f8a1b2c3d4e5f60718293a4b5c6d7e8f90a1b2c3d4e5f\""}},"Cache-Control":{"description":"Freshness directives for a catalog page or entry. `private` because the answer is decided per credential (the response also carries `Vary: Authorization, X-API-Key`), so a shared cache must not hold it, and `must-revalidate` so a stale copy is revalidated against the `ETag` rather than served on. Error responses on these routes are `private, no-store` instead.","required":false,"schema":{"type":"string","example":"private, max-age=300, must-revalidate"}}},"content":{"_mintlify/placeholder":{"schema":{"type":"object","description":""}}}},"2fa33a34-4b23-4d67-8f34-fe19c54fc1d7":{"description":"A Router request-level failure - the request never reached the model, or failed for a reason the model itself did not report. The body is `RouterErrorResponse` and the bucket is repeated on `X-Comfy-Error-Type`.","headers":{"X-Comfy-Error-Type":{"description":"Coarse, machine-readable bucket for the failure, set by Router on every error response. It carries the same value as `RouterErrorResponse.error_type`, and on the `422` it is the only machine-readable bucket, because that body is the FastAPI `detail[]` shape and has no `error_type` field of its own. A client can therefore branch on this header alone, before deciding which of the two Router error bodies it received.","required":true,"schema":{"type":"string","description":"Coarse, machine-readable bucket for a Router failure, mirrored on the `X-Comfy-Error-Type` response header so a caller can branch without parsing the body. The set is closed at twenty values: the six request-level buckets `invalid_input`, `content_policy_violation`, `provider_error`, `provider_timeout`, `insufficient_credits` and `model_not_found`, plus the transport-level `unauthorized`, `forbidden`, `concurrency_limit_exceeded`, `client_disconnected`, `internal_error`, `deadline_exceeded`, `not_enabled`, `service_unavailable`, `rate_limited`, `cancelled`, `queue_timeout`, `request_not_found`, `queue_backlog_full` and `result_expired`. Closed describes the set as documented today, not a bound that holds forever: the set is expected to grow, which is why this is deliberately a plain string and not an `enum`, so a client must treat an unrecognised value as `internal_error` rather than switch exhaustively over the list above and break on the next addition.","example":"invalid_input","x-comfy-error-types":[{"value":"invalid_input","tier":"request","meaning":"The request was rejected before it reached the model - a malformed body, a malformed or expired pagination cursor, an input the model's own schema does not accept, or an `Idempotency-Key` that cannot serve this request (already used for a different request - the method, the path and query, or the body differ - or already consumed by a call whose response cannot be replayed). Sent with `409` in the key cases and with `400`/`422` in the others; the status says which, and the key cases are the ones answered by using a new key rather than by editing the request."},{"value":"content_policy_violation","tier":"request","meaning":"The provider refused the request on content-policy grounds. The refusal is deterministic: re-sending the same input will be refused again."},{"value":"provider_error","tier":"request","meaning":"The partner provider reported a failure of its own, or returned a response Router could not interpret as a result."},{"value":"provider_timeout","tier":"request","meaning":"The partner provider did not answer within its deadline. This bucket is the provider timing out and never Router's own server deadline, which is reported as `deadline_exceeded` - the two share `504` and are separated because they name different causes: this one says the partner failed, that one says Comfy stopped holding the connection."},{"value":"insufficient_credits","tier":"request","meaning":"The calling workspace does not have enough credits to run the model."},{"value":"model_not_found","tier":"request","meaning":"The `{provider}/{model}` ID names no model Router can run: an unknown provider, an unenrolled id, one the catalog has retired past its sunset date, or a catalogued ID the provider does not currently serve for Comfy. A retired id's `detail` reads `\u003cid> was retired on \u003cYYYY-MM-DD>; use \u003csuccessor>` when a successor is recorded and not withheld, or just `\u003cid> was retired on \u003cYYYY-MM-DD>` when it is not, instead of a suggestion list; the successor named for a retired id is always a live catalog ID you can re-point the call at, and a successor is dropped from `detail` either because the table records none, or because the recorded one is currently withheld behind an embargo the caller cannot see. A model whose sunset is still in the future is not withdrawn: it stays listed by `GET /v2/models` and stays runnable until the date, so a listing taken today remains the answer to what runs today. Otherwise `detail` carries up to three suggestions drawn from the models the caller is entitled to see, except for a catalogued ID the provider does not currently serve for Comfy, whose response carries no suggestions."},{"value":"unauthorized","tier":"transport","meaning":"The request carried no usable credential."},{"value":"forbidden","tier":"transport","meaning":"The credential is valid but is not entitled to this model or this operation."},{"value":"concurrency_limit_exceeded","tier":"transport","meaning":"The workspace already has as many calls in flight as it is allowed; retry once one of them finishes. It carries one further condition on the run route, on a `409` rather than the `429` above: another call is already in flight for the `Idempotency-Key` this request presented. Re-send the same key after `Retry-After` seconds to collect that call's result."},{"value":"client_disconnected","tier":"transport","meaning":"The caller closed the connection before Router could return a result. It is logged rather than delivered - there is no socket left to write it to - and it is an attribution, not a billing outcome: a provider generation that completed is billed regardless of whether the caller received the response."},{"value":"internal_error","tier":"transport","meaning":"Router itself failed. It is also the value a client should treat any unrecognized bucket as, so a later addition to the set does not break a client generated before it."},{"value":"deadline_exceeded","tier":"transport","meaning":"Comfy stopped holding the connection at its own configured bound before an answer arrived. It shares `504` with `provider_timeout` and the pair says which side ran out of time; this one is Comfy's own bound, so nothing about the request was rejected and the same request may be retried. It says nothing about the charge: a provider generation that completed is billed regardless of whether the caller received the response. Retry it with the same `Idempotency-Key`: when the provider had already accepted the generation, the retry collects that generation rather than dispatching another, and a `Retry-After` on the `504` says when to ask."},{"value":"not_enabled","tier":"transport","meaning":"Comfy Router is not switched on for this caller yet. Nothing about the request is wrong and the model exists, which is why this is not `model_not_found`; it shares `403` with `forbidden` and is not the same thing, because `forbidden` is an entitlement decision about the caller while this is a state of the rollout. It is terminal: do not retry, and do not treat it as an outage. The one exception to \"about the caller\" is the queued submit, which also answers `not_enabled` for a model whose partner answers a generation directly as bytes: that model cannot yet be queued, so it is the model and not the caller that is refused, nothing is queued or charged, and the synchronous route `POST /v2/models/{provider}/{model}` runs it instead."},{"value":"service_unavailable","tier":"transport","meaning":"A service Comfy Router depends on is temporarily unavailable and the caller did nothing wrong. Retry it with backoff: it is the one bucket here whose condition clears on its own, without the caller changing the request and without a concurrency slot freeing, which is what distinguishes it from the other retryable answers (`concurrency_limit_exceeded`, `deadline_exceeded`). It is separate from `internal_error` - which is a `500` and means Router itself failed - so a client can tell \"come back shortly\" from \"this call is not going to work\"."},{"value":"rate_limited","tier":"transport","meaning":"The caller has spent an allowance measured over a window and must wait for that window to roll. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one clears the moment one of the caller's own in-flight calls finishes, so retrying in seconds is right, whereas nothing the caller does drains this one early. `detail` names the window, and `Retry-After` the seconds to wait."},{"value":"cancelled","tier":"transport","meaning":"A queued request was withdrawn — through the cancel route, or by an operator — before it produced a result; it is terminal, and it is not by itself a statement about the charge. Cancelling stops Comfy waiting and it does not always stop the partner working, so a request cancelled while it was still `IN_QUEUE` was never dispatched and cannot be charged, whereas one cancelled after it was admitted may still be charged — a partner generation that completes is charged whether or not anyone collected it, which is why the cancel route calls the ask a request rather than a guarantee. It is not `client_disconnected`: that one says nobody is listening any more while a generation may still be running and billable, whereas this says the request itself was withdrawn. A caller polling the status read sees `200` with this in the body, because the request ended and so the read succeeded. Collecting the result of such a request answers `409` in Router's error envelope: that read also worked, and what it found is a request whose terminal state - the caller's own decision - leaves nothing to return. It is deliberately not `410`, which on that route means the result aged out of its retention window, and not a `5xx`, which would report the caller's own cancellation as a Router fault and read to a polling client as worth retrying."},{"value":"queue_timeout","tier":"transport","meaning":"A queued request waited past its queue timeout without ever being admitted. Terminal, unbilled, and it never took a concurrency slot — the job never reached a provider. Deliberately not `deadline_exceeded`, which is the synchronous route's connection bound: a `deadline_exceeded` generation may be running and billable, whereas this one provably never started. It is returned under `504`, which it shares with `provider_timeout` and `deadline_exceeded` because a status can only say that a clock ran out; which clock - the partner's, Comfy's connection bound, or the queue's admission bound - is what `error_type` carries. The request itself is terminal, so submit a new one rather than re-reading this one."},{"value":"request_not_found","tier":"transport","meaning":"The `request_id` names no request of the caller's under this model. It is the second of the two conditions the queued reads' `404` covers; the first is the `{provider}/{model}` ID resolving to no partner model, which is `model_not_found` and carries fuzzy model suggestions. It also covers the right-id / wrong-model URL the path shape refuses, and it is deliberately indistinguishable from a request in another workspace, so a probe with a guessed id learns nothing. A request that has merely aged out of its retention window is `410`, not this."},{"value":"queue_backlog_full","tier":"transport","meaning":"The caller already has too many queued requests waiting to run, so this submit was refused. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one is the synchronous route's answer for too many calls in flight at once, whereas the queue accepts a submit at that limit and parks it, and this bucket is the separate bound on how many a caller may leave waiting so that parking cannot mean enqueuing without end. It clears as the caller's own queued requests finish, so retry once some of them complete."},{"value":"result_expired","tier":"transport","meaning":"The request existed and finished, but its result is no longer stored, so the collect has nothing to return. It is returned under `410` and is permanent for that id: re-reading will answer the same way, so do not retry it. It is deliberately not `request_not_found`, which would tell you the id was never valid and invite re-submitting a generation that has already run. Running the model again means submitting a new request. For a workspace under a zero-data-retention agreement, a queued request's result is not kept after it is delivered: it is removed shortly after the first collect, and an uncollected result is removed after a grace period once the request finishes, so store the result on your side when you first collect it."}]}},"X-Comfy-Request-Id":{"description":"Server-generated identifier for this call, present on every Router response - success, 4xx and 5xx alike, because an error response is exactly when a user needs an id to quote in a support request. The same value is written into the call's usage/audit event, which is what lets a complaint about a charge be joined to the charge itself instead of searched for by timestamp.","required":true,"schema":{"type":"string","format":"uuid","example":"6f1a1a6e-6a53-4a5f-9d3a-2b3b0a1f9c21"}}},"content":{"application/json":{"schema":{"type":"object","description":"Router's request-level error body: what is returned when the request never reached the model, or failed for a reason the model itself did not report - auth, quota, an unknown model ID, or provider transport. A model-level validation failure has its own shape, `RouterValidationErrorResponse`, because flattening a FastAPI `detail[]` array into this `detail` string would destroy the per-field granularity an SDK branches on.","properties":{"detail":{"type":"string","description":"Human-readable description of the failure, safe to surface to an end user. Not machine-parsed - branch on `error_type` instead. For `model_not_found` on a retired model, `detail` reads `\u003cid> was retired on \u003cYYYY-MM-DD>; use \u003csuccessor>` when a successor is recorded and visible to the caller, or just `\u003cid> was retired on \u003cYYYY-MM-DD>` when it is not, instead of the usual suggestion list. A successor named here is always a live catalog ID the call can be re-pointed at. A model whose sunset date is still in the future is not retired: it stays listed by `GET /v2/models` and stays runnable until that date. Like the rest of `detail`, this retirement wording is for display and carries no compatibility guarantee.","isRequired":true,"uniqueKey":"detail","typeLabel":"string"},"error_type":{"type":"string","description":"Coarse, machine-readable bucket for a Router failure, mirrored on the `X-Comfy-Error-Type` response header so a caller can branch without parsing the body. The set is closed at twenty values: the six request-level buckets `invalid_input`, `content_policy_violation`, `provider_error`, `provider_timeout`, `insufficient_credits` and `model_not_found`, plus the transport-level `unauthorized`, `forbidden`, `concurrency_limit_exceeded`, `client_disconnected`, `internal_error`, `deadline_exceeded`, `not_enabled`, `service_unavailable`, `rate_limited`, `cancelled`, `queue_timeout`, `request_not_found`, `queue_backlog_full` and `result_expired`. Closed describes the set as documented today, not a bound that holds forever: the set is expected to grow, which is why this is deliberately a plain string and not an `enum`, so a client must treat an unrecognised value as `internal_error` rather than switch exhaustively over the list above and break on the next addition.","example":"invalid_input","x-comfy-error-types":[{"value":"invalid_input","tier":"request","meaning":"The request was rejected before it reached the model - a malformed body, a malformed or expired pagination cursor, an input the model's own schema does not accept, or an `Idempotency-Key` that cannot serve this request (already used for a different request - the method, the path and query, or the body differ - or already consumed by a call whose response cannot be replayed). Sent with `409` in the key cases and with `400`/`422` in the others; the status says which, and the key cases are the ones answered by using a new key rather than by editing the request."},{"value":"content_policy_violation","tier":"request","meaning":"The provider refused the request on content-policy grounds. The refusal is deterministic: re-sending the same input will be refused again."},{"value":"provider_error","tier":"request","meaning":"The partner provider reported a failure of its own, or returned a response Router could not interpret as a result."},{"value":"provider_timeout","tier":"request","meaning":"The partner provider did not answer within its deadline. This bucket is the provider timing out and never Router's own server deadline, which is reported as `deadline_exceeded` - the two share `504` and are separated because they name different causes: this one says the partner failed, that one says Comfy stopped holding the connection."},{"value":"insufficient_credits","tier":"request","meaning":"The calling workspace does not have enough credits to run the model."},{"value":"model_not_found","tier":"request","meaning":"The `{provider}/{model}` ID names no model Router can run: an unknown provider, an unenrolled id, one the catalog has retired past its sunset date, or a catalogued ID the provider does not currently serve for Comfy. A retired id's `detail` reads `\u003cid> was retired on \u003cYYYY-MM-DD>; use \u003csuccessor>` when a successor is recorded and not withheld, or just `\u003cid> was retired on \u003cYYYY-MM-DD>` when it is not, instead of a suggestion list; the successor named for a retired id is always a live catalog ID you can re-point the call at, and a successor is dropped from `detail` either because the table records none, or because the recorded one is currently withheld behind an embargo the caller cannot see. A model whose sunset is still in the future is not withdrawn: it stays listed by `GET /v2/models` and stays runnable until the date, so a listing taken today remains the answer to what runs today. Otherwise `detail` carries up to three suggestions drawn from the models the caller is entitled to see, except for a catalogued ID the provider does not currently serve for Comfy, whose response carries no suggestions."},{"value":"unauthorized","tier":"transport","meaning":"The request carried no usable credential."},{"value":"forbidden","tier":"transport","meaning":"The credential is valid but is not entitled to this model or this operation."},{"value":"concurrency_limit_exceeded","tier":"transport","meaning":"The workspace already has as many calls in flight as it is allowed; retry once one of them finishes. It carries one further condition on the run route, on a `409` rather than the `429` above: another call is already in flight for the `Idempotency-Key` this request presented. Re-send the same key after `Retry-After` seconds to collect that call's result."},{"value":"client_disconnected","tier":"transport","meaning":"The caller closed the connection before Router could return a result. It is logged rather than delivered - there is no socket left to write it to - and it is an attribution, not a billing outcome: a provider generation that completed is billed regardless of whether the caller received the response."},{"value":"internal_error","tier":"transport","meaning":"Router itself failed. It is also the value a client should treat any unrecognized bucket as, so a later addition to the set does not break a client generated before it."},{"value":"deadline_exceeded","tier":"transport","meaning":"Comfy stopped holding the connection at its own configured bound before an answer arrived. It shares `504` with `provider_timeout` and the pair says which side ran out of time; this one is Comfy's own bound, so nothing about the request was rejected and the same request may be retried. It says nothing about the charge: a provider generation that completed is billed regardless of whether the caller received the response. Retry it with the same `Idempotency-Key`: when the provider had already accepted the generation, the retry collects that generation rather than dispatching another, and a `Retry-After` on the `504` says when to ask."},{"value":"not_enabled","tier":"transport","meaning":"Comfy Router is not switched on for this caller yet. Nothing about the request is wrong and the model exists, which is why this is not `model_not_found`; it shares `403` with `forbidden` and is not the same thing, because `forbidden` is an entitlement decision about the caller while this is a state of the rollout. It is terminal: do not retry, and do not treat it as an outage. The one exception to \"about the caller\" is the queued submit, which also answers `not_enabled` for a model whose partner answers a generation directly as bytes: that model cannot yet be queued, so it is the model and not the caller that is refused, nothing is queued or charged, and the synchronous route `POST /v2/models/{provider}/{model}` runs it instead."},{"value":"service_unavailable","tier":"transport","meaning":"A service Comfy Router depends on is temporarily unavailable and the caller did nothing wrong. Retry it with backoff: it is the one bucket here whose condition clears on its own, without the caller changing the request and without a concurrency slot freeing, which is what distinguishes it from the other retryable answers (`concurrency_limit_exceeded`, `deadline_exceeded`). It is separate from `internal_error` - which is a `500` and means Router itself failed - so a client can tell \"come back shortly\" from \"this call is not going to work\"."},{"value":"rate_limited","tier":"transport","meaning":"The caller has spent an allowance measured over a window and must wait for that window to roll. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one clears the moment one of the caller's own in-flight calls finishes, so retrying in seconds is right, whereas nothing the caller does drains this one early. `detail` names the window, and `Retry-After` the seconds to wait."},{"value":"cancelled","tier":"transport","meaning":"A queued request was withdrawn — through the cancel route, or by an operator — before it produced a result; it is terminal, and it is not by itself a statement about the charge. Cancelling stops Comfy waiting and it does not always stop the partner working, so a request cancelled while it was still `IN_QUEUE` was never dispatched and cannot be charged, whereas one cancelled after it was admitted may still be charged — a partner generation that completes is charged whether or not anyone collected it, which is why the cancel route calls the ask a request rather than a guarantee. It is not `client_disconnected`: that one says nobody is listening any more while a generation may still be running and billable, whereas this says the request itself was withdrawn. A caller polling the status read sees `200` with this in the body, because the request ended and so the read succeeded. Collecting the result of such a request answers `409` in Router's error envelope: that read also worked, and what it found is a request whose terminal state - the caller's own decision - leaves nothing to return. It is deliberately not `410`, which on that route means the result aged out of its retention window, and not a `5xx`, which would report the caller's own cancellation as a Router fault and read to a polling client as worth retrying."},{"value":"queue_timeout","tier":"transport","meaning":"A queued request waited past its queue timeout without ever being admitted. Terminal, unbilled, and it never took a concurrency slot — the job never reached a provider. Deliberately not `deadline_exceeded`, which is the synchronous route's connection bound: a `deadline_exceeded` generation may be running and billable, whereas this one provably never started. It is returned under `504`, which it shares with `provider_timeout` and `deadline_exceeded` because a status can only say that a clock ran out; which clock - the partner's, Comfy's connection bound, or the queue's admission bound - is what `error_type` carries. The request itself is terminal, so submit a new one rather than re-reading this one."},{"value":"request_not_found","tier":"transport","meaning":"The `request_id` names no request of the caller's under this model. It is the second of the two conditions the queued reads' `404` covers; the first is the `{provider}/{model}` ID resolving to no partner model, which is `model_not_found` and carries fuzzy model suggestions. It also covers the right-id / wrong-model URL the path shape refuses, and it is deliberately indistinguishable from a request in another workspace, so a probe with a guessed id learns nothing. A request that has merely aged out of its retention window is `410`, not this."},{"value":"queue_backlog_full","tier":"transport","meaning":"The caller already has too many queued requests waiting to run, so this submit was refused. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one is the synchronous route's answer for too many calls in flight at once, whereas the queue accepts a submit at that limit and parks it, and this bucket is the separate bound on how many a caller may leave waiting so that parking cannot mean enqueuing without end. It clears as the caller's own queued requests finish, so retry once some of them complete."},{"value":"result_expired","tier":"transport","meaning":"The request existed and finished, but its result is no longer stored, so the collect has nothing to return. It is returned under `410` and is permanent for that id: re-reading will answer the same way, so do not retry it. It is deliberately not `request_not_found`, which would tell you the id was never valid and invite re-submitting a generation that has already run. Running the model again means submitting a new request. For a workspace under a zero-data-retention agreement, a queued request's result is not kept after it is delivered: it is removed shortly after the first collect, and an uncollected result is removed after a grace period once the request finishes, so store the result on your side when you first collect it."}],"isRequired":true,"uniqueKey":"error_type","typeLabel":"string"},"upstream_detail":{"type":"string","description":"A bounded, sanitized reason the model provider gave for rejecting the request, present only when `error_type` is `invalid_input` and `X-Comfy-Upstream-Status` is a provider `4xx` or `2xx` - i.e. the provider refused the request as malformed and said why. Usually that status is a `4xx`; it is a `2xx` for a provider that reports a rejected generation inside a success envelope (a BytePlus failed-task poll is HTTP `200` with the reason in its body). Absent on every other failure, including provider `5xx`, transport failures, content-policy refusals and any refusal Router raised about itself. It mirrors the `X-Comfy-Upstream-Detail` header. The queued collect (`GET /v2/models/{provider}/{model}/requests/{request_id}`) of a run that ended `invalid_input` carries the same reason in this field on its first collect, which then deletes it from the request: a later collect normally answers the same `error_type` and `detail` without it, so read it from the first collect. That read sends neither header.","uniqueKey":"upstream_detail","typeLabel":"string"},"refusal_subject":{"type":"string","description":"Which input or output a content-policy refusal was about, as a Router-level closed vocabulary: `input`, `output`, `input_text`, `input_image`, `input_video`, `input_audio`, `output_text`, `output_image`, `output_video`, `output_audio`. The bare `input` / `output` values name the side when the provider did not name a modality. Present only when `error_type` is `content_policy_violation` and the provider named the refused subject with a machine-readable code; absent otherwise. Never provider text. Named today for BytePlus, Runway, BFL, Gemini, Veo, Vertex, xAI and Wan refusals, and for the MiniMax Router's Nano Banana legs' Gemini refusals; a provider whose refusal does not say which side it was about leaves it absent. It mirrors the `X-Comfy-Refusal-Subject` header.","uniqueKey":"refusal_subject","typeLabel":"string"},"denied_reason":{"type":"string","description":"Billing's policy branch for a partner-node spend denial - `no_funds` when the billed wallet is empty, `unspecified` when the call was refused without a funds reading saying so. Present only when `error_type` is `insufficient_credits` and the deployment reports wallet facts; absent otherwise. A machine-readable code, not display text: show `detail`, branch on this.","uniqueKey":"denied_reason","typeLabel":"string"},"billed_workspace":{"type":"object","description":"The wallet a refused call would have charged, on `RouterErrorResponse.billed_workspace`. Answers \"which wallet is empty?\" - it is not necessarily the workspace the caller's UI currently has selected, and a team member's personal credential names their personal wallet here even while a funded team is open. Present only alongside `denied_reason`.","properties":{"type":{"type":"string","description":"`personal` or `team`. An open string: treat any other value as personal.","isRequired":true,"uniqueKey":"billed_workspace.type","typeLabel":"string"},"id":{"type":"string","description":"The Cloud workspace id. Absent on the legacy rail.","uniqueKey":"billed_workspace.id","typeLabel":"string"},"rail":{"type":"string","description":"`legacy` when the call bills the caller's pre-workspace Metronome customer rather than a Cloud workspace; that customer has no workspace id. Absent for a Cloud workspace.","uniqueKey":"billed_workspace.rail","typeLabel":"string"}},"required":["type"],"uniqueKey":"billed_workspace","typeLabel":"object"},"credential_kind":{"type":"string","description":"How the refused request authenticated - `personal_api_key`, `team_api_key`, `personal_session` or `team_session`. Answers \"which credential do I switch to bill the other wallet?\". It is not a permission and grants nothing. Present only alongside `denied_reason`.","uniqueKey":"credential_kind","typeLabel":"string"}},"required":["detail","error_type"],"uniqueKey":"","typeLabel":"object"},"examples":{"Example":{"title":"Example","value":{"detail":"\u003cstring>","error_type":"invalid_input","upstream_detail":"\u003cstring>","refusal_subject":"\u003cstring>","denied_reason":"\u003cstring>","billed_workspace":{"type":"\u003cstring>","id":"\u003cstring>","rail":"\u003cstring>"},"credential_kind":"\u003cstring>"}}},"exampleType":"generated"}},"exampleType":"generated"},"91a10445-9225-4d4d-a974-7c7bf264501a":{"description":"A Router request-level failure - the request never reached the model, or failed for a reason the model itself did not report. The body is `RouterErrorResponse` and the bucket is repeated on `X-Comfy-Error-Type`.","headers":{"X-Comfy-Error-Type":{"description":"Coarse, machine-readable bucket for the failure, set by Router on every error response. It carries the same value as `RouterErrorResponse.error_type`, and on the `422` it is the only machine-readable bucket, because that body is the FastAPI `detail[]` shape and has no `error_type` field of its own. A client can therefore branch on this header alone, before deciding which of the two Router error bodies it received.","required":true,"schema":{"type":"string","description":"Coarse, machine-readable bucket for a Router failure, mirrored on the `X-Comfy-Error-Type` response header so a caller can branch without parsing the body. The set is closed at twenty values: the six request-level buckets `invalid_input`, `content_policy_violation`, `provider_error`, `provider_timeout`, `insufficient_credits` and `model_not_found`, plus the transport-level `unauthorized`, `forbidden`, `concurrency_limit_exceeded`, `client_disconnected`, `internal_error`, `deadline_exceeded`, `not_enabled`, `service_unavailable`, `rate_limited`, `cancelled`, `queue_timeout`, `request_not_found`, `queue_backlog_full` and `result_expired`. Closed describes the set as documented today, not a bound that holds forever: the set is expected to grow, which is why this is deliberately a plain string and not an `enum`, so a client must treat an unrecognised value as `internal_error` rather than switch exhaustively over the list above and break on the next addition.","example":"invalid_input","x-comfy-error-types":[{"value":"invalid_input","tier":"request","meaning":"The request was rejected before it reached the model - a malformed body, a malformed or expired pagination cursor, an input the model's own schema does not accept, or an `Idempotency-Key` that cannot serve this request (already used for a different request - the method, the path and query, or the body differ - or already consumed by a call whose response cannot be replayed). Sent with `409` in the key cases and with `400`/`422` in the others; the status says which, and the key cases are the ones answered by using a new key rather than by editing the request."},{"value":"content_policy_violation","tier":"request","meaning":"The provider refused the request on content-policy grounds. The refusal is deterministic: re-sending the same input will be refused again."},{"value":"provider_error","tier":"request","meaning":"The partner provider reported a failure of its own, or returned a response Router could not interpret as a result."},{"value":"provider_timeout","tier":"request","meaning":"The partner provider did not answer within its deadline. This bucket is the provider timing out and never Router's own server deadline, which is reported as `deadline_exceeded` - the two share `504` and are separated because they name different causes: this one says the partner failed, that one says Comfy stopped holding the connection."},{"value":"insufficient_credits","tier":"request","meaning":"The calling workspace does not have enough credits to run the model."},{"value":"model_not_found","tier":"request","meaning":"The `{provider}/{model}` ID names no model Router can run: an unknown provider, an unenrolled id, one the catalog has retired past its sunset date, or a catalogued ID the provider does not currently serve for Comfy. A retired id's `detail` reads `\u003cid> was retired on \u003cYYYY-MM-DD>; use \u003csuccessor>` when a successor is recorded and not withheld, or just `\u003cid> was retired on \u003cYYYY-MM-DD>` when it is not, instead of a suggestion list; the successor named for a retired id is always a live catalog ID you can re-point the call at, and a successor is dropped from `detail` either because the table records none, or because the recorded one is currently withheld behind an embargo the caller cannot see. A model whose sunset is still in the future is not withdrawn: it stays listed by `GET /v2/models` and stays runnable until the date, so a listing taken today remains the answer to what runs today. Otherwise `detail` carries up to three suggestions drawn from the models the caller is entitled to see, except for a catalogued ID the provider does not currently serve for Comfy, whose response carries no suggestions."},{"value":"unauthorized","tier":"transport","meaning":"The request carried no usable credential."},{"value":"forbidden","tier":"transport","meaning":"The credential is valid but is not entitled to this model or this operation."},{"value":"concurrency_limit_exceeded","tier":"transport","meaning":"The workspace already has as many calls in flight as it is allowed; retry once one of them finishes. It carries one further condition on the run route, on a `409` rather than the `429` above: another call is already in flight for the `Idempotency-Key` this request presented. Re-send the same key after `Retry-After` seconds to collect that call's result."},{"value":"client_disconnected","tier":"transport","meaning":"The caller closed the connection before Router could return a result. It is logged rather than delivered - there is no socket left to write it to - and it is an attribution, not a billing outcome: a provider generation that completed is billed regardless of whether the caller received the response."},{"value":"internal_error","tier":"transport","meaning":"Router itself failed. It is also the value a client should treat any unrecognized bucket as, so a later addition to the set does not break a client generated before it."},{"value":"deadline_exceeded","tier":"transport","meaning":"Comfy stopped holding the connection at its own configured bound before an answer arrived. It shares `504` with `provider_timeout` and the pair says which side ran out of time; this one is Comfy's own bound, so nothing about the request was rejected and the same request may be retried. It says nothing about the charge: a provider generation that completed is billed regardless of whether the caller received the response. Retry it with the same `Idempotency-Key`: when the provider had already accepted the generation, the retry collects that generation rather than dispatching another, and a `Retry-After` on the `504` says when to ask."},{"value":"not_enabled","tier":"transport","meaning":"Comfy Router is not switched on for this caller yet. Nothing about the request is wrong and the model exists, which is why this is not `model_not_found`; it shares `403` with `forbidden` and is not the same thing, because `forbidden` is an entitlement decision about the caller while this is a state of the rollout. It is terminal: do not retry, and do not treat it as an outage. The one exception to \"about the caller\" is the queued submit, which also answers `not_enabled` for a model whose partner answers a generation directly as bytes: that model cannot yet be queued, so it is the model and not the caller that is refused, nothing is queued or charged, and the synchronous route `POST /v2/models/{provider}/{model}` runs it instead."},{"value":"service_unavailable","tier":"transport","meaning":"A service Comfy Router depends on is temporarily unavailable and the caller did nothing wrong. Retry it with backoff: it is the one bucket here whose condition clears on its own, without the caller changing the request and without a concurrency slot freeing, which is what distinguishes it from the other retryable answers (`concurrency_limit_exceeded`, `deadline_exceeded`). It is separate from `internal_error` - which is a `500` and means Router itself failed - so a client can tell \"come back shortly\" from \"this call is not going to work\"."},{"value":"rate_limited","tier":"transport","meaning":"The caller has spent an allowance measured over a window and must wait for that window to roll. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one clears the moment one of the caller's own in-flight calls finishes, so retrying in seconds is right, whereas nothing the caller does drains this one early. `detail` names the window, and `Retry-After` the seconds to wait."},{"value":"cancelled","tier":"transport","meaning":"A queued request was withdrawn — through the cancel route, or by an operator — before it produced a result; it is terminal, and it is not by itself a statement about the charge. Cancelling stops Comfy waiting and it does not always stop the partner working, so a request cancelled while it was still `IN_QUEUE` was never dispatched and cannot be charged, whereas one cancelled after it was admitted may still be charged — a partner generation that completes is charged whether or not anyone collected it, which is why the cancel route calls the ask a request rather than a guarantee. It is not `client_disconnected`: that one says nobody is listening any more while a generation may still be running and billable, whereas this says the request itself was withdrawn. A caller polling the status read sees `200` with this in the body, because the request ended and so the read succeeded. Collecting the result of such a request answers `409` in Router's error envelope: that read also worked, and what it found is a request whose terminal state - the caller's own decision - leaves nothing to return. It is deliberately not `410`, which on that route means the result aged out of its retention window, and not a `5xx`, which would report the caller's own cancellation as a Router fault and read to a polling client as worth retrying."},{"value":"queue_timeout","tier":"transport","meaning":"A queued request waited past its queue timeout without ever being admitted. Terminal, unbilled, and it never took a concurrency slot — the job never reached a provider. Deliberately not `deadline_exceeded`, which is the synchronous route's connection bound: a `deadline_exceeded` generation may be running and billable, whereas this one provably never started. It is returned under `504`, which it shares with `provider_timeout` and `deadline_exceeded` because a status can only say that a clock ran out; which clock - the partner's, Comfy's connection bound, or the queue's admission bound - is what `error_type` carries. The request itself is terminal, so submit a new one rather than re-reading this one."},{"value":"request_not_found","tier":"transport","meaning":"The `request_id` names no request of the caller's under this model. It is the second of the two conditions the queued reads' `404` covers; the first is the `{provider}/{model}` ID resolving to no partner model, which is `model_not_found` and carries fuzzy model suggestions. It also covers the right-id / wrong-model URL the path shape refuses, and it is deliberately indistinguishable from a request in another workspace, so a probe with a guessed id learns nothing. A request that has merely aged out of its retention window is `410`, not this."},{"value":"queue_backlog_full","tier":"transport","meaning":"The caller already has too many queued requests waiting to run, so this submit was refused. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one is the synchronous route's answer for too many calls in flight at once, whereas the queue accepts a submit at that limit and parks it, and this bucket is the separate bound on how many a caller may leave waiting so that parking cannot mean enqueuing without end. It clears as the caller's own queued requests finish, so retry once some of them complete."},{"value":"result_expired","tier":"transport","meaning":"The request existed and finished, but its result is no longer stored, so the collect has nothing to return. It is returned under `410` and is permanent for that id: re-reading will answer the same way, so do not retry it. It is deliberately not `request_not_found`, which would tell you the id was never valid and invite re-submitting a generation that has already run. Running the model again means submitting a new request. For a workspace under a zero-data-retention agreement, a queued request's result is not kept after it is delivered: it is removed shortly after the first collect, and an uncollected result is removed after a grace period once the request finishes, so store the result on your side when you first collect it."}]}},"X-Comfy-Request-Id":{"description":"Server-generated identifier for this call, present on every Router response - success, 4xx and 5xx alike, because an error response is exactly when a user needs an id to quote in a support request. The same value is written into the call's usage/audit event, which is what lets a complaint about a charge be joined to the charge itself instead of searched for by timestamp.","required":true,"schema":{"type":"string","format":"uuid","example":"6f1a1a6e-6a53-4a5f-9d3a-2b3b0a1f9c21"}}},"content":{"application/json":{"schema":{"type":"object","description":"Router's request-level error body: what is returned when the request never reached the model, or failed for a reason the model itself did not report - auth, quota, an unknown model ID, or provider transport. A model-level validation failure has its own shape, `RouterValidationErrorResponse`, because flattening a FastAPI `detail[]` array into this `detail` string would destroy the per-field granularity an SDK branches on.","properties":{"detail":{"type":"string","description":"Human-readable description of the failure, safe to surface to an end user. Not machine-parsed - branch on `error_type` instead. For `model_not_found` on a retired model, `detail` reads `\u003cid> was retired on \u003cYYYY-MM-DD>; use \u003csuccessor>` when a successor is recorded and visible to the caller, or just `\u003cid> was retired on \u003cYYYY-MM-DD>` when it is not, instead of the usual suggestion list. A successor named here is always a live catalog ID the call can be re-pointed at. A model whose sunset date is still in the future is not retired: it stays listed by `GET /v2/models` and stays runnable until that date. Like the rest of `detail`, this retirement wording is for display and carries no compatibility guarantee.","isRequired":true,"uniqueKey":"detail","typeLabel":"string"},"error_type":{"type":"string","description":"Coarse, machine-readable bucket for a Router failure, mirrored on the `X-Comfy-Error-Type` response header so a caller can branch without parsing the body. The set is closed at twenty values: the six request-level buckets `invalid_input`, `content_policy_violation`, `provider_error`, `provider_timeout`, `insufficient_credits` and `model_not_found`, plus the transport-level `unauthorized`, `forbidden`, `concurrency_limit_exceeded`, `client_disconnected`, `internal_error`, `deadline_exceeded`, `not_enabled`, `service_unavailable`, `rate_limited`, `cancelled`, `queue_timeout`, `request_not_found`, `queue_backlog_full` and `result_expired`. Closed describes the set as documented today, not a bound that holds forever: the set is expected to grow, which is why this is deliberately a plain string and not an `enum`, so a client must treat an unrecognised value as `internal_error` rather than switch exhaustively over the list above and break on the next addition.","example":"invalid_input","x-comfy-error-types":[{"value":"invalid_input","tier":"request","meaning":"The request was rejected before it reached the model - a malformed body, a malformed or expired pagination cursor, an input the model's own schema does not accept, or an `Idempotency-Key` that cannot serve this request (already used for a different request - the method, the path and query, or the body differ - or already consumed by a call whose response cannot be replayed). Sent with `409` in the key cases and with `400`/`422` in the others; the status says which, and the key cases are the ones answered by using a new key rather than by editing the request."},{"value":"content_policy_violation","tier":"request","meaning":"The provider refused the request on content-policy grounds. The refusal is deterministic: re-sending the same input will be refused again."},{"value":"provider_error","tier":"request","meaning":"The partner provider reported a failure of its own, or returned a response Router could not interpret as a result."},{"value":"provider_timeout","tier":"request","meaning":"The partner provider did not answer within its deadline. This bucket is the provider timing out and never Router's own server deadline, which is reported as `deadline_exceeded` - the two share `504` and are separated because they name different causes: this one says the partner failed, that one says Comfy stopped holding the connection."},{"value":"insufficient_credits","tier":"request","meaning":"The calling workspace does not have enough credits to run the model."},{"value":"model_not_found","tier":"request","meaning":"The `{provider}/{model}` ID names no model Router can run: an unknown provider, an unenrolled id, one the catalog has retired past its sunset date, or a catalogued ID the provider does not currently serve for Comfy. A retired id's `detail` reads `\u003cid> was retired on \u003cYYYY-MM-DD>; use \u003csuccessor>` when a successor is recorded and not withheld, or just `\u003cid> was retired on \u003cYYYY-MM-DD>` when it is not, instead of a suggestion list; the successor named for a retired id is always a live catalog ID you can re-point the call at, and a successor is dropped from `detail` either because the table records none, or because the recorded one is currently withheld behind an embargo the caller cannot see. A model whose sunset is still in the future is not withdrawn: it stays listed by `GET /v2/models` and stays runnable until the date, so a listing taken today remains the answer to what runs today. Otherwise `detail` carries up to three suggestions drawn from the models the caller is entitled to see, except for a catalogued ID the provider does not currently serve for Comfy, whose response carries no suggestions."},{"value":"unauthorized","tier":"transport","meaning":"The request carried no usable credential."},{"value":"forbidden","tier":"transport","meaning":"The credential is valid but is not entitled to this model or this operation."},{"value":"concurrency_limit_exceeded","tier":"transport","meaning":"The workspace already has as many calls in flight as it is allowed; retry once one of them finishes. It carries one further condition on the run route, on a `409` rather than the `429` above: another call is already in flight for the `Idempotency-Key` this request presented. Re-send the same key after `Retry-After` seconds to collect that call's result."},{"value":"client_disconnected","tier":"transport","meaning":"The caller closed the connection before Router could return a result. It is logged rather than delivered - there is no socket left to write it to - and it is an attribution, not a billing outcome: a provider generation that completed is billed regardless of whether the caller received the response."},{"value":"internal_error","tier":"transport","meaning":"Router itself failed. It is also the value a client should treat any unrecognized bucket as, so a later addition to the set does not break a client generated before it."},{"value":"deadline_exceeded","tier":"transport","meaning":"Comfy stopped holding the connection at its own configured bound before an answer arrived. It shares `504` with `provider_timeout` and the pair says which side ran out of time; this one is Comfy's own bound, so nothing about the request was rejected and the same request may be retried. It says nothing about the charge: a provider generation that completed is billed regardless of whether the caller received the response. Retry it with the same `Idempotency-Key`: when the provider had already accepted the generation, the retry collects that generation rather than dispatching another, and a `Retry-After` on the `504` says when to ask."},{"value":"not_enabled","tier":"transport","meaning":"Comfy Router is not switched on for this caller yet. Nothing about the request is wrong and the model exists, which is why this is not `model_not_found`; it shares `403` with `forbidden` and is not the same thing, because `forbidden` is an entitlement decision about the caller while this is a state of the rollout. It is terminal: do not retry, and do not treat it as an outage. The one exception to \"about the caller\" is the queued submit, which also answers `not_enabled` for a model whose partner answers a generation directly as bytes: that model cannot yet be queued, so it is the model and not the caller that is refused, nothing is queued or charged, and the synchronous route `POST /v2/models/{provider}/{model}` runs it instead."},{"value":"service_unavailable","tier":"transport","meaning":"A service Comfy Router depends on is temporarily unavailable and the caller did nothing wrong. Retry it with backoff: it is the one bucket here whose condition clears on its own, without the caller changing the request and without a concurrency slot freeing, which is what distinguishes it from the other retryable answers (`concurrency_limit_exceeded`, `deadline_exceeded`). It is separate from `internal_error` - which is a `500` and means Router itself failed - so a client can tell \"come back shortly\" from \"this call is not going to work\"."},{"value":"rate_limited","tier":"transport","meaning":"The caller has spent an allowance measured over a window and must wait for that window to roll. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one clears the moment one of the caller's own in-flight calls finishes, so retrying in seconds is right, whereas nothing the caller does drains this one early. `detail` names the window, and `Retry-After` the seconds to wait."},{"value":"cancelled","tier":"transport","meaning":"A queued request was withdrawn — through the cancel route, or by an operator — before it produced a result; it is terminal, and it is not by itself a statement about the charge. Cancelling stops Comfy waiting and it does not always stop the partner working, so a request cancelled while it was still `IN_QUEUE` was never dispatched and cannot be charged, whereas one cancelled after it was admitted may still be charged — a partner generation that completes is charged whether or not anyone collected it, which is why the cancel route calls the ask a request rather than a guarantee. It is not `client_disconnected`: that one says nobody is listening any more while a generation may still be running and billable, whereas this says the request itself was withdrawn. A caller polling the status read sees `200` with this in the body, because the request ended and so the read succeeded. Collecting the result of such a request answers `409` in Router's error envelope: that read also worked, and what it found is a request whose terminal state - the caller's own decision - leaves nothing to return. It is deliberately not `410`, which on that route means the result aged out of its retention window, and not a `5xx`, which would report the caller's own cancellation as a Router fault and read to a polling client as worth retrying."},{"value":"queue_timeout","tier":"transport","meaning":"A queued request waited past its queue timeout without ever being admitted. Terminal, unbilled, and it never took a concurrency slot — the job never reached a provider. Deliberately not `deadline_exceeded`, which is the synchronous route's connection bound: a `deadline_exceeded` generation may be running and billable, whereas this one provably never started. It is returned under `504`, which it shares with `provider_timeout` and `deadline_exceeded` because a status can only say that a clock ran out; which clock - the partner's, Comfy's connection bound, or the queue's admission bound - is what `error_type` carries. The request itself is terminal, so submit a new one rather than re-reading this one."},{"value":"request_not_found","tier":"transport","meaning":"The `request_id` names no request of the caller's under this model. It is the second of the two conditions the queued reads' `404` covers; the first is the `{provider}/{model}` ID resolving to no partner model, which is `model_not_found` and carries fuzzy model suggestions. It also covers the right-id / wrong-model URL the path shape refuses, and it is deliberately indistinguishable from a request in another workspace, so a probe with a guessed id learns nothing. A request that has merely aged out of its retention window is `410`, not this."},{"value":"queue_backlog_full","tier":"transport","meaning":"The caller already has too many queued requests waiting to run, so this submit was refused. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one is the synchronous route's answer for too many calls in flight at once, whereas the queue accepts a submit at that limit and parks it, and this bucket is the separate bound on how many a caller may leave waiting so that parking cannot mean enqueuing without end. It clears as the caller's own queued requests finish, so retry once some of them complete."},{"value":"result_expired","tier":"transport","meaning":"The request existed and finished, but its result is no longer stored, so the collect has nothing to return. It is returned under `410` and is permanent for that id: re-reading will answer the same way, so do not retry it. It is deliberately not `request_not_found`, which would tell you the id was never valid and invite re-submitting a generation that has already run. Running the model again means submitting a new request. For a workspace under a zero-data-retention agreement, a queued request's result is not kept after it is delivered: it is removed shortly after the first collect, and an uncollected result is removed after a grace period once the request finishes, so store the result on your side when you first collect it."}],"isRequired":true,"uniqueKey":"error_type","typeLabel":"string"},"upstream_detail":{"type":"string","description":"A bounded, sanitized reason the model provider gave for rejecting the request, present only when `error_type` is `invalid_input` and `X-Comfy-Upstream-Status` is a provider `4xx` or `2xx` - i.e. the provider refused the request as malformed and said why. Usually that status is a `4xx`; it is a `2xx` for a provider that reports a rejected generation inside a success envelope (a BytePlus failed-task poll is HTTP `200` with the reason in its body). Absent on every other failure, including provider `5xx`, transport failures, content-policy refusals and any refusal Router raised about itself. It mirrors the `X-Comfy-Upstream-Detail` header. The queued collect (`GET /v2/models/{provider}/{model}/requests/{request_id}`) of a run that ended `invalid_input` carries the same reason in this field on its first collect, which then deletes it from the request: a later collect normally answers the same `error_type` and `detail` without it, so read it from the first collect. That read sends neither header.","uniqueKey":"upstream_detail","typeLabel":"string"},"refusal_subject":{"type":"string","description":"Which input or output a content-policy refusal was about, as a Router-level closed vocabulary: `input`, `output`, `input_text`, `input_image`, `input_video`, `input_audio`, `output_text`, `output_image`, `output_video`, `output_audio`. The bare `input` / `output` values name the side when the provider did not name a modality. Present only when `error_type` is `content_policy_violation` and the provider named the refused subject with a machine-readable code; absent otherwise. Never provider text. Named today for BytePlus, Runway, BFL, Gemini, Veo, Vertex, xAI and Wan refusals, and for the MiniMax Router's Nano Banana legs' Gemini refusals; a provider whose refusal does not say which side it was about leaves it absent. It mirrors the `X-Comfy-Refusal-Subject` header.","uniqueKey":"refusal_subject","typeLabel":"string"},"denied_reason":{"type":"string","description":"Billing's policy branch for a partner-node spend denial - `no_funds` when the billed wallet is empty, `unspecified` when the call was refused without a funds reading saying so. Present only when `error_type` is `insufficient_credits` and the deployment reports wallet facts; absent otherwise. A machine-readable code, not display text: show `detail`, branch on this.","uniqueKey":"denied_reason","typeLabel":"string"},"billed_workspace":{"type":"object","description":"The wallet a refused call would have charged, on `RouterErrorResponse.billed_workspace`. Answers \"which wallet is empty?\" - it is not necessarily the workspace the caller's UI currently has selected, and a team member's personal credential names their personal wallet here even while a funded team is open. Present only alongside `denied_reason`.","properties":{"type":{"type":"string","description":"`personal` or `team`. An open string: treat any other value as personal.","isRequired":true,"uniqueKey":"billed_workspace.type","typeLabel":"string"},"id":{"type":"string","description":"The Cloud workspace id. Absent on the legacy rail.","uniqueKey":"billed_workspace.id","typeLabel":"string"},"rail":{"type":"string","description":"`legacy` when the call bills the caller's pre-workspace Metronome customer rather than a Cloud workspace; that customer has no workspace id. Absent for a Cloud workspace.","uniqueKey":"billed_workspace.rail","typeLabel":"string"}},"required":["type"],"uniqueKey":"billed_workspace","typeLabel":"object"},"credential_kind":{"type":"string","description":"How the refused request authenticated - `personal_api_key`, `team_api_key`, `personal_session` or `team_session`. Answers \"which credential do I switch to bill the other wallet?\". It is not a permission and grants nothing. Present only alongside `denied_reason`.","uniqueKey":"credential_kind","typeLabel":"string"}},"required":["detail","error_type"],"uniqueKey":"","typeLabel":"object"},"examples":{"Example":{"title":"Example","value":{"detail":"\u003cstring>","error_type":"invalid_input","upstream_detail":"\u003cstring>","refusal_subject":"\u003cstring>","denied_reason":"\u003cstring>","billed_workspace":{"type":"\u003cstring>","id":"\u003cstring>","rail":"\u003cstring>"},"credential_kind":"\u003cstring>"}}},"exampleType":"generated"}},"exampleType":"generated"},"3171abdf-95c0-44af-b054-6eac9e52f872":{"description":"A Router request-level failure - the request never reached the model, or failed for a reason the model itself did not report. The body is `RouterErrorResponse` and the bucket is repeated on `X-Comfy-Error-Type`.","headers":{"X-Comfy-Error-Type":{"description":"Coarse, machine-readable bucket for the failure, set by Router on every error response. It carries the same value as `RouterErrorResponse.error_type`, and on the `422` it is the only machine-readable bucket, because that body is the FastAPI `detail[]` shape and has no `error_type` field of its own. A client can therefore branch on this header alone, before deciding which of the two Router error bodies it received.","required":true,"schema":{"type":"string","description":"Coarse, machine-readable bucket for a Router failure, mirrored on the `X-Comfy-Error-Type` response header so a caller can branch without parsing the body. The set is closed at twenty values: the six request-level buckets `invalid_input`, `content_policy_violation`, `provider_error`, `provider_timeout`, `insufficient_credits` and `model_not_found`, plus the transport-level `unauthorized`, `forbidden`, `concurrency_limit_exceeded`, `client_disconnected`, `internal_error`, `deadline_exceeded`, `not_enabled`, `service_unavailable`, `rate_limited`, `cancelled`, `queue_timeout`, `request_not_found`, `queue_backlog_full` and `result_expired`. Closed describes the set as documented today, not a bound that holds forever: the set is expected to grow, which is why this is deliberately a plain string and not an `enum`, so a client must treat an unrecognised value as `internal_error` rather than switch exhaustively over the list above and break on the next addition.","example":"invalid_input","x-comfy-error-types":[{"value":"invalid_input","tier":"request","meaning":"The request was rejected before it reached the model - a malformed body, a malformed or expired pagination cursor, an input the model's own schema does not accept, or an `Idempotency-Key` that cannot serve this request (already used for a different request - the method, the path and query, or the body differ - or already consumed by a call whose response cannot be replayed). Sent with `409` in the key cases and with `400`/`422` in the others; the status says which, and the key cases are the ones answered by using a new key rather than by editing the request."},{"value":"content_policy_violation","tier":"request","meaning":"The provider refused the request on content-policy grounds. The refusal is deterministic: re-sending the same input will be refused again."},{"value":"provider_error","tier":"request","meaning":"The partner provider reported a failure of its own, or returned a response Router could not interpret as a result."},{"value":"provider_timeout","tier":"request","meaning":"The partner provider did not answer within its deadline. This bucket is the provider timing out and never Router's own server deadline, which is reported as `deadline_exceeded` - the two share `504` and are separated because they name different causes: this one says the partner failed, that one says Comfy stopped holding the connection."},{"value":"insufficient_credits","tier":"request","meaning":"The calling workspace does not have enough credits to run the model."},{"value":"model_not_found","tier":"request","meaning":"The `{provider}/{model}` ID names no model Router can run: an unknown provider, an unenrolled id, one the catalog has retired past its sunset date, or a catalogued ID the provider does not currently serve for Comfy. A retired id's `detail` reads `\u003cid> was retired on \u003cYYYY-MM-DD>; use \u003csuccessor>` when a successor is recorded and not withheld, or just `\u003cid> was retired on \u003cYYYY-MM-DD>` when it is not, instead of a suggestion list; the successor named for a retired id is always a live catalog ID you can re-point the call at, and a successor is dropped from `detail` either because the table records none, or because the recorded one is currently withheld behind an embargo the caller cannot see. A model whose sunset is still in the future is not withdrawn: it stays listed by `GET /v2/models` and stays runnable until the date, so a listing taken today remains the answer to what runs today. Otherwise `detail` carries up to three suggestions drawn from the models the caller is entitled to see, except for a catalogued ID the provider does not currently serve for Comfy, whose response carries no suggestions."},{"value":"unauthorized","tier":"transport","meaning":"The request carried no usable credential."},{"value":"forbidden","tier":"transport","meaning":"The credential is valid but is not entitled to this model or this operation."},{"value":"concurrency_limit_exceeded","tier":"transport","meaning":"The workspace already has as many calls in flight as it is allowed; retry once one of them finishes. It carries one further condition on the run route, on a `409` rather than the `429` above: another call is already in flight for the `Idempotency-Key` this request presented. Re-send the same key after `Retry-After` seconds to collect that call's result."},{"value":"client_disconnected","tier":"transport","meaning":"The caller closed the connection before Router could return a result. It is logged rather than delivered - there is no socket left to write it to - and it is an attribution, not a billing outcome: a provider generation that completed is billed regardless of whether the caller received the response."},{"value":"internal_error","tier":"transport","meaning":"Router itself failed. It is also the value a client should treat any unrecognized bucket as, so a later addition to the set does not break a client generated before it."},{"value":"deadline_exceeded","tier":"transport","meaning":"Comfy stopped holding the connection at its own configured bound before an answer arrived. It shares `504` with `provider_timeout` and the pair says which side ran out of time; this one is Comfy's own bound, so nothing about the request was rejected and the same request may be retried. It says nothing about the charge: a provider generation that completed is billed regardless of whether the caller received the response. Retry it with the same `Idempotency-Key`: when the provider had already accepted the generation, the retry collects that generation rather than dispatching another, and a `Retry-After` on the `504` says when to ask."},{"value":"not_enabled","tier":"transport","meaning":"Comfy Router is not switched on for this caller yet. Nothing about the request is wrong and the model exists, which is why this is not `model_not_found`; it shares `403` with `forbidden` and is not the same thing, because `forbidden` is an entitlement decision about the caller while this is a state of the rollout. It is terminal: do not retry, and do not treat it as an outage. The one exception to \"about the caller\" is the queued submit, which also answers `not_enabled` for a model whose partner answers a generation directly as bytes: that model cannot yet be queued, so it is the model and not the caller that is refused, nothing is queued or charged, and the synchronous route `POST /v2/models/{provider}/{model}` runs it instead."},{"value":"service_unavailable","tier":"transport","meaning":"A service Comfy Router depends on is temporarily unavailable and the caller did nothing wrong. Retry it with backoff: it is the one bucket here whose condition clears on its own, without the caller changing the request and without a concurrency slot freeing, which is what distinguishes it from the other retryable answers (`concurrency_limit_exceeded`, `deadline_exceeded`). It is separate from `internal_error` - which is a `500` and means Router itself failed - so a client can tell \"come back shortly\" from \"this call is not going to work\"."},{"value":"rate_limited","tier":"transport","meaning":"The caller has spent an allowance measured over a window and must wait for that window to roll. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one clears the moment one of the caller's own in-flight calls finishes, so retrying in seconds is right, whereas nothing the caller does drains this one early. `detail` names the window, and `Retry-After` the seconds to wait."},{"value":"cancelled","tier":"transport","meaning":"A queued request was withdrawn — through the cancel route, or by an operator — before it produced a result; it is terminal, and it is not by itself a statement about the charge. Cancelling stops Comfy waiting and it does not always stop the partner working, so a request cancelled while it was still `IN_QUEUE` was never dispatched and cannot be charged, whereas one cancelled after it was admitted may still be charged — a partner generation that completes is charged whether or not anyone collected it, which is why the cancel route calls the ask a request rather than a guarantee. It is not `client_disconnected`: that one says nobody is listening any more while a generation may still be running and billable, whereas this says the request itself was withdrawn. A caller polling the status read sees `200` with this in the body, because the request ended and so the read succeeded. Collecting the result of such a request answers `409` in Router's error envelope: that read also worked, and what it found is a request whose terminal state - the caller's own decision - leaves nothing to return. It is deliberately not `410`, which on that route means the result aged out of its retention window, and not a `5xx`, which would report the caller's own cancellation as a Router fault and read to a polling client as worth retrying."},{"value":"queue_timeout","tier":"transport","meaning":"A queued request waited past its queue timeout without ever being admitted. Terminal, unbilled, and it never took a concurrency slot — the job never reached a provider. Deliberately not `deadline_exceeded`, which is the synchronous route's connection bound: a `deadline_exceeded` generation may be running and billable, whereas this one provably never started. It is returned under `504`, which it shares with `provider_timeout` and `deadline_exceeded` because a status can only say that a clock ran out; which clock - the partner's, Comfy's connection bound, or the queue's admission bound - is what `error_type` carries. The request itself is terminal, so submit a new one rather than re-reading this one."},{"value":"request_not_found","tier":"transport","meaning":"The `request_id` names no request of the caller's under this model. It is the second of the two conditions the queued reads' `404` covers; the first is the `{provider}/{model}` ID resolving to no partner model, which is `model_not_found` and carries fuzzy model suggestions. It also covers the right-id / wrong-model URL the path shape refuses, and it is deliberately indistinguishable from a request in another workspace, so a probe with a guessed id learns nothing. A request that has merely aged out of its retention window is `410`, not this."},{"value":"queue_backlog_full","tier":"transport","meaning":"The caller already has too many queued requests waiting to run, so this submit was refused. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one is the synchronous route's answer for too many calls in flight at once, whereas the queue accepts a submit at that limit and parks it, and this bucket is the separate bound on how many a caller may leave waiting so that parking cannot mean enqueuing without end. It clears as the caller's own queued requests finish, so retry once some of them complete."},{"value":"result_expired","tier":"transport","meaning":"The request existed and finished, but its result is no longer stored, so the collect has nothing to return. It is returned under `410` and is permanent for that id: re-reading will answer the same way, so do not retry it. It is deliberately not `request_not_found`, which would tell you the id was never valid and invite re-submitting a generation that has already run. Running the model again means submitting a new request. For a workspace under a zero-data-retention agreement, a queued request's result is not kept after it is delivered: it is removed shortly after the first collect, and an uncollected result is removed after a grace period once the request finishes, so store the result on your side when you first collect it."}]}},"X-Comfy-Request-Id":{"description":"Server-generated identifier for this call, present on every Router response - success, 4xx and 5xx alike, because an error response is exactly when a user needs an id to quote in a support request. The same value is written into the call's usage/audit event, which is what lets a complaint about a charge be joined to the charge itself instead of searched for by timestamp.","required":true,"schema":{"type":"string","format":"uuid","example":"6f1a1a6e-6a53-4a5f-9d3a-2b3b0a1f9c21"}}},"content":{"application/json":{"schema":{"type":"object","description":"Router's request-level error body: what is returned when the request never reached the model, or failed for a reason the model itself did not report - auth, quota, an unknown model ID, or provider transport. A model-level validation failure has its own shape, `RouterValidationErrorResponse`, because flattening a FastAPI `detail[]` array into this `detail` string would destroy the per-field granularity an SDK branches on.","properties":{"detail":{"type":"string","description":"Human-readable description of the failure, safe to surface to an end user. Not machine-parsed - branch on `error_type` instead. For `model_not_found` on a retired model, `detail` reads `\u003cid> was retired on \u003cYYYY-MM-DD>; use \u003csuccessor>` when a successor is recorded and visible to the caller, or just `\u003cid> was retired on \u003cYYYY-MM-DD>` when it is not, instead of the usual suggestion list. A successor named here is always a live catalog ID the call can be re-pointed at. A model whose sunset date is still in the future is not retired: it stays listed by `GET /v2/models` and stays runnable until that date. Like the rest of `detail`, this retirement wording is for display and carries no compatibility guarantee.","isRequired":true,"uniqueKey":"detail","typeLabel":"string"},"error_type":{"type":"string","description":"Coarse, machine-readable bucket for a Router failure, mirrored on the `X-Comfy-Error-Type` response header so a caller can branch without parsing the body. The set is closed at twenty values: the six request-level buckets `invalid_input`, `content_policy_violation`, `provider_error`, `provider_timeout`, `insufficient_credits` and `model_not_found`, plus the transport-level `unauthorized`, `forbidden`, `concurrency_limit_exceeded`, `client_disconnected`, `internal_error`, `deadline_exceeded`, `not_enabled`, `service_unavailable`, `rate_limited`, `cancelled`, `queue_timeout`, `request_not_found`, `queue_backlog_full` and `result_expired`. Closed describes the set as documented today, not a bound that holds forever: the set is expected to grow, which is why this is deliberately a plain string and not an `enum`, so a client must treat an unrecognised value as `internal_error` rather than switch exhaustively over the list above and break on the next addition.","example":"invalid_input","x-comfy-error-types":[{"value":"invalid_input","tier":"request","meaning":"The request was rejected before it reached the model - a malformed body, a malformed or expired pagination cursor, an input the model's own schema does not accept, or an `Idempotency-Key` that cannot serve this request (already used for a different request - the method, the path and query, or the body differ - or already consumed by a call whose response cannot be replayed). Sent with `409` in the key cases and with `400`/`422` in the others; the status says which, and the key cases are the ones answered by using a new key rather than by editing the request."},{"value":"content_policy_violation","tier":"request","meaning":"The provider refused the request on content-policy grounds. The refusal is deterministic: re-sending the same input will be refused again."},{"value":"provider_error","tier":"request","meaning":"The partner provider reported a failure of its own, or returned a response Router could not interpret as a result."},{"value":"provider_timeout","tier":"request","meaning":"The partner provider did not answer within its deadline. This bucket is the provider timing out and never Router's own server deadline, which is reported as `deadline_exceeded` - the two share `504` and are separated because they name different causes: this one says the partner failed, that one says Comfy stopped holding the connection."},{"value":"insufficient_credits","tier":"request","meaning":"The calling workspace does not have enough credits to run the model."},{"value":"model_not_found","tier":"request","meaning":"The `{provider}/{model}` ID names no model Router can run: an unknown provider, an unenrolled id, one the catalog has retired past its sunset date, or a catalogued ID the provider does not currently serve for Comfy. A retired id's `detail` reads `\u003cid> was retired on \u003cYYYY-MM-DD>; use \u003csuccessor>` when a successor is recorded and not withheld, or just `\u003cid> was retired on \u003cYYYY-MM-DD>` when it is not, instead of a suggestion list; the successor named for a retired id is always a live catalog ID you can re-point the call at, and a successor is dropped from `detail` either because the table records none, or because the recorded one is currently withheld behind an embargo the caller cannot see. A model whose sunset is still in the future is not withdrawn: it stays listed by `GET /v2/models` and stays runnable until the date, so a listing taken today remains the answer to what runs today. Otherwise `detail` carries up to three suggestions drawn from the models the caller is entitled to see, except for a catalogued ID the provider does not currently serve for Comfy, whose response carries no suggestions."},{"value":"unauthorized","tier":"transport","meaning":"The request carried no usable credential."},{"value":"forbidden","tier":"transport","meaning":"The credential is valid but is not entitled to this model or this operation."},{"value":"concurrency_limit_exceeded","tier":"transport","meaning":"The workspace already has as many calls in flight as it is allowed; retry once one of them finishes. It carries one further condition on the run route, on a `409` rather than the `429` above: another call is already in flight for the `Idempotency-Key` this request presented. Re-send the same key after `Retry-After` seconds to collect that call's result."},{"value":"client_disconnected","tier":"transport","meaning":"The caller closed the connection before Router could return a result. It is logged rather than delivered - there is no socket left to write it to - and it is an attribution, not a billing outcome: a provider generation that completed is billed regardless of whether the caller received the response."},{"value":"internal_error","tier":"transport","meaning":"Router itself failed. It is also the value a client should treat any unrecognized bucket as, so a later addition to the set does not break a client generated before it."},{"value":"deadline_exceeded","tier":"transport","meaning":"Comfy stopped holding the connection at its own configured bound before an answer arrived. It shares `504` with `provider_timeout` and the pair says which side ran out of time; this one is Comfy's own bound, so nothing about the request was rejected and the same request may be retried. It says nothing about the charge: a provider generation that completed is billed regardless of whether the caller received the response. Retry it with the same `Idempotency-Key`: when the provider had already accepted the generation, the retry collects that generation rather than dispatching another, and a `Retry-After` on the `504` says when to ask."},{"value":"not_enabled","tier":"transport","meaning":"Comfy Router is not switched on for this caller yet. Nothing about the request is wrong and the model exists, which is why this is not `model_not_found`; it shares `403` with `forbidden` and is not the same thing, because `forbidden` is an entitlement decision about the caller while this is a state of the rollout. It is terminal: do not retry, and do not treat it as an outage. The one exception to \"about the caller\" is the queued submit, which also answers `not_enabled` for a model whose partner answers a generation directly as bytes: that model cannot yet be queued, so it is the model and not the caller that is refused, nothing is queued or charged, and the synchronous route `POST /v2/models/{provider}/{model}` runs it instead."},{"value":"service_unavailable","tier":"transport","meaning":"A service Comfy Router depends on is temporarily unavailable and the caller did nothing wrong. Retry it with backoff: it is the one bucket here whose condition clears on its own, without the caller changing the request and without a concurrency slot freeing, which is what distinguishes it from the other retryable answers (`concurrency_limit_exceeded`, `deadline_exceeded`). It is separate from `internal_error` - which is a `500` and means Router itself failed - so a client can tell \"come back shortly\" from \"this call is not going to work\"."},{"value":"rate_limited","tier":"transport","meaning":"The caller has spent an allowance measured over a window and must wait for that window to roll. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one clears the moment one of the caller's own in-flight calls finishes, so retrying in seconds is right, whereas nothing the caller does drains this one early. `detail` names the window, and `Retry-After` the seconds to wait."},{"value":"cancelled","tier":"transport","meaning":"A queued request was withdrawn — through the cancel route, or by an operator — before it produced a result; it is terminal, and it is not by itself a statement about the charge. Cancelling stops Comfy waiting and it does not always stop the partner working, so a request cancelled while it was still `IN_QUEUE` was never dispatched and cannot be charged, whereas one cancelled after it was admitted may still be charged — a partner generation that completes is charged whether or not anyone collected it, which is why the cancel route calls the ask a request rather than a guarantee. It is not `client_disconnected`: that one says nobody is listening any more while a generation may still be running and billable, whereas this says the request itself was withdrawn. A caller polling the status read sees `200` with this in the body, because the request ended and so the read succeeded. Collecting the result of such a request answers `409` in Router's error envelope: that read also worked, and what it found is a request whose terminal state - the caller's own decision - leaves nothing to return. It is deliberately not `410`, which on that route means the result aged out of its retention window, and not a `5xx`, which would report the caller's own cancellation as a Router fault and read to a polling client as worth retrying."},{"value":"queue_timeout","tier":"transport","meaning":"A queued request waited past its queue timeout without ever being admitted. Terminal, unbilled, and it never took a concurrency slot — the job never reached a provider. Deliberately not `deadline_exceeded`, which is the synchronous route's connection bound: a `deadline_exceeded` generation may be running and billable, whereas this one provably never started. It is returned under `504`, which it shares with `provider_timeout` and `deadline_exceeded` because a status can only say that a clock ran out; which clock - the partner's, Comfy's connection bound, or the queue's admission bound - is what `error_type` carries. The request itself is terminal, so submit a new one rather than re-reading this one."},{"value":"request_not_found","tier":"transport","meaning":"The `request_id` names no request of the caller's under this model. It is the second of the two conditions the queued reads' `404` covers; the first is the `{provider}/{model}` ID resolving to no partner model, which is `model_not_found` and carries fuzzy model suggestions. It also covers the right-id / wrong-model URL the path shape refuses, and it is deliberately indistinguishable from a request in another workspace, so a probe with a guessed id learns nothing. A request that has merely aged out of its retention window is `410`, not this."},{"value":"queue_backlog_full","tier":"transport","meaning":"The caller already has too many queued requests waiting to run, so this submit was refused. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one is the synchronous route's answer for too many calls in flight at once, whereas the queue accepts a submit at that limit and parks it, and this bucket is the separate bound on how many a caller may leave waiting so that parking cannot mean enqueuing without end. It clears as the caller's own queued requests finish, so retry once some of them complete."},{"value":"result_expired","tier":"transport","meaning":"The request existed and finished, but its result is no longer stored, so the collect has nothing to return. It is returned under `410` and is permanent for that id: re-reading will answer the same way, so do not retry it. It is deliberately not `request_not_found`, which would tell you the id was never valid and invite re-submitting a generation that has already run. Running the model again means submitting a new request. For a workspace under a zero-data-retention agreement, a queued request's result is not kept after it is delivered: it is removed shortly after the first collect, and an uncollected result is removed after a grace period once the request finishes, so store the result on your side when you first collect it."}],"isRequired":true,"uniqueKey":"error_type","typeLabel":"string"},"upstream_detail":{"type":"string","description":"A bounded, sanitized reason the model provider gave for rejecting the request, present only when `error_type` is `invalid_input` and `X-Comfy-Upstream-Status` is a provider `4xx` or `2xx` - i.e. the provider refused the request as malformed and said why. Usually that status is a `4xx`; it is a `2xx` for a provider that reports a rejected generation inside a success envelope (a BytePlus failed-task poll is HTTP `200` with the reason in its body). Absent on every other failure, including provider `5xx`, transport failures, content-policy refusals and any refusal Router raised about itself. It mirrors the `X-Comfy-Upstream-Detail` header. The queued collect (`GET /v2/models/{provider}/{model}/requests/{request_id}`) of a run that ended `invalid_input` carries the same reason in this field on its first collect, which then deletes it from the request: a later collect normally answers the same `error_type` and `detail` without it, so read it from the first collect. That read sends neither header.","uniqueKey":"upstream_detail","typeLabel":"string"},"refusal_subject":{"type":"string","description":"Which input or output a content-policy refusal was about, as a Router-level closed vocabulary: `input`, `output`, `input_text`, `input_image`, `input_video`, `input_audio`, `output_text`, `output_image`, `output_video`, `output_audio`. The bare `input` / `output` values name the side when the provider did not name a modality. Present only when `error_type` is `content_policy_violation` and the provider named the refused subject with a machine-readable code; absent otherwise. Never provider text. Named today for BytePlus, Runway, BFL, Gemini, Veo, Vertex, xAI and Wan refusals, and for the MiniMax Router's Nano Banana legs' Gemini refusals; a provider whose refusal does not say which side it was about leaves it absent. It mirrors the `X-Comfy-Refusal-Subject` header.","uniqueKey":"refusal_subject","typeLabel":"string"},"denied_reason":{"type":"string","description":"Billing's policy branch for a partner-node spend denial - `no_funds` when the billed wallet is empty, `unspecified` when the call was refused without a funds reading saying so. Present only when `error_type` is `insufficient_credits` and the deployment reports wallet facts; absent otherwise. A machine-readable code, not display text: show `detail`, branch on this.","uniqueKey":"denied_reason","typeLabel":"string"},"billed_workspace":{"type":"object","description":"The wallet a refused call would have charged, on `RouterErrorResponse.billed_workspace`. Answers \"which wallet is empty?\" - it is not necessarily the workspace the caller's UI currently has selected, and a team member's personal credential names their personal wallet here even while a funded team is open. Present only alongside `denied_reason`.","properties":{"type":{"type":"string","description":"`personal` or `team`. An open string: treat any other value as personal.","isRequired":true,"uniqueKey":"billed_workspace.type","typeLabel":"string"},"id":{"type":"string","description":"The Cloud workspace id. Absent on the legacy rail.","uniqueKey":"billed_workspace.id","typeLabel":"string"},"rail":{"type":"string","description":"`legacy` when the call bills the caller's pre-workspace Metronome customer rather than a Cloud workspace; that customer has no workspace id. Absent for a Cloud workspace.","uniqueKey":"billed_workspace.rail","typeLabel":"string"}},"required":["type"],"uniqueKey":"billed_workspace","typeLabel":"object"},"credential_kind":{"type":"string","description":"How the refused request authenticated - `personal_api_key`, `team_api_key`, `personal_session` or `team_session`. Answers \"which credential do I switch to bill the other wallet?\". It is not a permission and grants nothing. Present only alongside `denied_reason`.","uniqueKey":"credential_kind","typeLabel":"string"}},"required":["detail","error_type"],"uniqueKey":"","typeLabel":"object"},"examples":{"Example":{"title":"Example","value":{"detail":"\u003cstring>","error_type":"invalid_input","upstream_detail":"\u003cstring>","refusal_subject":"\u003cstring>","denied_reason":"\u003cstring>","billed_workspace":{"type":"\u003cstring>","id":"\u003cstring>","rail":"\u003cstring>"},"credential_kind":"\u003cstring>"}}},"exampleType":"generated"}},"exampleType":"generated"},"7dfa64fa-4be2-4272-8d6a-a542b27c997b":{"description":"The caller has spent their per-caller request allowance, which refills continuously over a one-minute window, and the request was refused before Router did anything with it. The bucket is always `rate_limited`: wait the number of seconds in `Retry-After` before sending the next request. Nothing was charged, and a request sent sooner is refused again. The body is `RouterErrorResponse` and the bucket is repeated on `X-Comfy-Error-Type`.","headers":{"X-Comfy-Error-Type":{"description":"Coarse, machine-readable bucket for the failure, set by Router on every error response. It carries the same value as `RouterErrorResponse.error_type`, and on the `422` it is the only machine-readable bucket, because that body is the FastAPI `detail[]` shape and has no `error_type` field of its own. A client can therefore branch on this header alone, before deciding which of the two Router error bodies it received.","required":true,"schema":{"type":"string","description":"Coarse, machine-readable bucket for a Router failure, mirrored on the `X-Comfy-Error-Type` response header so a caller can branch without parsing the body. The set is closed at twenty values: the six request-level buckets `invalid_input`, `content_policy_violation`, `provider_error`, `provider_timeout`, `insufficient_credits` and `model_not_found`, plus the transport-level `unauthorized`, `forbidden`, `concurrency_limit_exceeded`, `client_disconnected`, `internal_error`, `deadline_exceeded`, `not_enabled`, `service_unavailable`, `rate_limited`, `cancelled`, `queue_timeout`, `request_not_found`, `queue_backlog_full` and `result_expired`. Closed describes the set as documented today, not a bound that holds forever: the set is expected to grow, which is why this is deliberately a plain string and not an `enum`, so a client must treat an unrecognised value as `internal_error` rather than switch exhaustively over the list above and break on the next addition.","example":"invalid_input","x-comfy-error-types":[{"value":"invalid_input","tier":"request","meaning":"The request was rejected before it reached the model - a malformed body, a malformed or expired pagination cursor, an input the model's own schema does not accept, or an `Idempotency-Key` that cannot serve this request (already used for a different request - the method, the path and query, or the body differ - or already consumed by a call whose response cannot be replayed). Sent with `409` in the key cases and with `400`/`422` in the others; the status says which, and the key cases are the ones answered by using a new key rather than by editing the request."},{"value":"content_policy_violation","tier":"request","meaning":"The provider refused the request on content-policy grounds. The refusal is deterministic: re-sending the same input will be refused again."},{"value":"provider_error","tier":"request","meaning":"The partner provider reported a failure of its own, or returned a response Router could not interpret as a result."},{"value":"provider_timeout","tier":"request","meaning":"The partner provider did not answer within its deadline. This bucket is the provider timing out and never Router's own server deadline, which is reported as `deadline_exceeded` - the two share `504` and are separated because they name different causes: this one says the partner failed, that one says Comfy stopped holding the connection."},{"value":"insufficient_credits","tier":"request","meaning":"The calling workspace does not have enough credits to run the model."},{"value":"model_not_found","tier":"request","meaning":"The `{provider}/{model}` ID names no model Router can run: an unknown provider, an unenrolled id, one the catalog has retired past its sunset date, or a catalogued ID the provider does not currently serve for Comfy. A retired id's `detail` reads `\u003cid> was retired on \u003cYYYY-MM-DD>; use \u003csuccessor>` when a successor is recorded and not withheld, or just `\u003cid> was retired on \u003cYYYY-MM-DD>` when it is not, instead of a suggestion list; the successor named for a retired id is always a live catalog ID you can re-point the call at, and a successor is dropped from `detail` either because the table records none, or because the recorded one is currently withheld behind an embargo the caller cannot see. A model whose sunset is still in the future is not withdrawn: it stays listed by `GET /v2/models` and stays runnable until the date, so a listing taken today remains the answer to what runs today. Otherwise `detail` carries up to three suggestions drawn from the models the caller is entitled to see, except for a catalogued ID the provider does not currently serve for Comfy, whose response carries no suggestions."},{"value":"unauthorized","tier":"transport","meaning":"The request carried no usable credential."},{"value":"forbidden","tier":"transport","meaning":"The credential is valid but is not entitled to this model or this operation."},{"value":"concurrency_limit_exceeded","tier":"transport","meaning":"The workspace already has as many calls in flight as it is allowed; retry once one of them finishes. It carries one further condition on the run route, on a `409` rather than the `429` above: another call is already in flight for the `Idempotency-Key` this request presented. Re-send the same key after `Retry-After` seconds to collect that call's result."},{"value":"client_disconnected","tier":"transport","meaning":"The caller closed the connection before Router could return a result. It is logged rather than delivered - there is no socket left to write it to - and it is an attribution, not a billing outcome: a provider generation that completed is billed regardless of whether the caller received the response."},{"value":"internal_error","tier":"transport","meaning":"Router itself failed. It is also the value a client should treat any unrecognized bucket as, so a later addition to the set does not break a client generated before it."},{"value":"deadline_exceeded","tier":"transport","meaning":"Comfy stopped holding the connection at its own configured bound before an answer arrived. It shares `504` with `provider_timeout` and the pair says which side ran out of time; this one is Comfy's own bound, so nothing about the request was rejected and the same request may be retried. It says nothing about the charge: a provider generation that completed is billed regardless of whether the caller received the response. Retry it with the same `Idempotency-Key`: when the provider had already accepted the generation, the retry collects that generation rather than dispatching another, and a `Retry-After` on the `504` says when to ask."},{"value":"not_enabled","tier":"transport","meaning":"Comfy Router is not switched on for this caller yet. Nothing about the request is wrong and the model exists, which is why this is not `model_not_found`; it shares `403` with `forbidden` and is not the same thing, because `forbidden` is an entitlement decision about the caller while this is a state of the rollout. It is terminal: do not retry, and do not treat it as an outage. The one exception to \"about the caller\" is the queued submit, which also answers `not_enabled` for a model whose partner answers a generation directly as bytes: that model cannot yet be queued, so it is the model and not the caller that is refused, nothing is queued or charged, and the synchronous route `POST /v2/models/{provider}/{model}` runs it instead."},{"value":"service_unavailable","tier":"transport","meaning":"A service Comfy Router depends on is temporarily unavailable and the caller did nothing wrong. Retry it with backoff: it is the one bucket here whose condition clears on its own, without the caller changing the request and without a concurrency slot freeing, which is what distinguishes it from the other retryable answers (`concurrency_limit_exceeded`, `deadline_exceeded`). It is separate from `internal_error` - which is a `500` and means Router itself failed - so a client can tell \"come back shortly\" from \"this call is not going to work\"."},{"value":"rate_limited","tier":"transport","meaning":"The caller has spent an allowance measured over a window and must wait for that window to roll. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one clears the moment one of the caller's own in-flight calls finishes, so retrying in seconds is right, whereas nothing the caller does drains this one early. `detail` names the window, and `Retry-After` the seconds to wait."},{"value":"cancelled","tier":"transport","meaning":"A queued request was withdrawn — through the cancel route, or by an operator — before it produced a result; it is terminal, and it is not by itself a statement about the charge. Cancelling stops Comfy waiting and it does not always stop the partner working, so a request cancelled while it was still `IN_QUEUE` was never dispatched and cannot be charged, whereas one cancelled after it was admitted may still be charged — a partner generation that completes is charged whether or not anyone collected it, which is why the cancel route calls the ask a request rather than a guarantee. It is not `client_disconnected`: that one says nobody is listening any more while a generation may still be running and billable, whereas this says the request itself was withdrawn. A caller polling the status read sees `200` with this in the body, because the request ended and so the read succeeded. Collecting the result of such a request answers `409` in Router's error envelope: that read also worked, and what it found is a request whose terminal state - the caller's own decision - leaves nothing to return. It is deliberately not `410`, which on that route means the result aged out of its retention window, and not a `5xx`, which would report the caller's own cancellation as a Router fault and read to a polling client as worth retrying."},{"value":"queue_timeout","tier":"transport","meaning":"A queued request waited past its queue timeout without ever being admitted. Terminal, unbilled, and it never took a concurrency slot — the job never reached a provider. Deliberately not `deadline_exceeded`, which is the synchronous route's connection bound: a `deadline_exceeded` generation may be running and billable, whereas this one provably never started. It is returned under `504`, which it shares with `provider_timeout` and `deadline_exceeded` because a status can only say that a clock ran out; which clock - the partner's, Comfy's connection bound, or the queue's admission bound - is what `error_type` carries. The request itself is terminal, so submit a new one rather than re-reading this one."},{"value":"request_not_found","tier":"transport","meaning":"The `request_id` names no request of the caller's under this model. It is the second of the two conditions the queued reads' `404` covers; the first is the `{provider}/{model}` ID resolving to no partner model, which is `model_not_found` and carries fuzzy model suggestions. It also covers the right-id / wrong-model URL the path shape refuses, and it is deliberately indistinguishable from a request in another workspace, so a probe with a guessed id learns nothing. A request that has merely aged out of its retention window is `410`, not this."},{"value":"queue_backlog_full","tier":"transport","meaning":"The caller already has too many queued requests waiting to run, so this submit was refused. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one is the synchronous route's answer for too many calls in flight at once, whereas the queue accepts a submit at that limit and parks it, and this bucket is the separate bound on how many a caller may leave waiting so that parking cannot mean enqueuing without end. It clears as the caller's own queued requests finish, so retry once some of them complete."},{"value":"result_expired","tier":"transport","meaning":"The request existed and finished, but its result is no longer stored, so the collect has nothing to return. It is returned under `410` and is permanent for that id: re-reading will answer the same way, so do not retry it. It is deliberately not `request_not_found`, which would tell you the id was never valid and invite re-submitting a generation that has already run. Running the model again means submitting a new request. For a workspace under a zero-data-retention agreement, a queued request's result is not kept after it is delivered: it is removed shortly after the first collect, and an uncollected result is removed after a grace period once the request finishes, so store the result on your side when you first collect it."}]}},"X-Comfy-Request-Id":{"description":"Server-generated identifier for this call, present on every Router response - success, 4xx and 5xx alike, because an error response is exactly when a user needs an id to quote in a support request. The same value is written into the call's usage/audit event, which is what lets a complaint about a charge be joined to the charge itself instead of searched for by timestamp.","required":true,"schema":{"type":"string","format":"uuid","example":"6f1a1a6e-6a53-4a5f-9d3a-2b3b0a1f9c21"}},"Retry-After":{"description":"Seconds to wait before sending another request after a `429` carrying `error_type: rate_limited`: the time until the caller's per-minute request allowance has refilled enough to admit one more request. It is sent on every `rate_limited` refusal and on no other answer, so it is absent on a `429` whose `X-Comfy-Error-Type` is `concurrency_limit_exceeded` or `queue_backlog_full`. Nothing the caller does drains it early; a request sent sooner is refused again and spends nothing, but gains nothing either.","required":false,"schema":{"type":"integer","minimum":1,"example":2}}},"content":{"application/json":{"schema":{"type":"object","description":"Router's request-level error body: what is returned when the request never reached the model, or failed for a reason the model itself did not report - auth, quota, an unknown model ID, or provider transport. A model-level validation failure has its own shape, `RouterValidationErrorResponse`, because flattening a FastAPI `detail[]` array into this `detail` string would destroy the per-field granularity an SDK branches on.","properties":{"detail":{"type":"string","description":"Human-readable description of the failure, safe to surface to an end user. Not machine-parsed - branch on `error_type` instead. For `model_not_found` on a retired model, `detail` reads `\u003cid> was retired on \u003cYYYY-MM-DD>; use \u003csuccessor>` when a successor is recorded and visible to the caller, or just `\u003cid> was retired on \u003cYYYY-MM-DD>` when it is not, instead of the usual suggestion list. A successor named here is always a live catalog ID the call can be re-pointed at. A model whose sunset date is still in the future is not retired: it stays listed by `GET /v2/models` and stays runnable until that date. Like the rest of `detail`, this retirement wording is for display and carries no compatibility guarantee.","isRequired":true,"uniqueKey":"detail","typeLabel":"string"},"error_type":{"type":"string","description":"Coarse, machine-readable bucket for a Router failure, mirrored on the `X-Comfy-Error-Type` response header so a caller can branch without parsing the body. The set is closed at twenty values: the six request-level buckets `invalid_input`, `content_policy_violation`, `provider_error`, `provider_timeout`, `insufficient_credits` and `model_not_found`, plus the transport-level `unauthorized`, `forbidden`, `concurrency_limit_exceeded`, `client_disconnected`, `internal_error`, `deadline_exceeded`, `not_enabled`, `service_unavailable`, `rate_limited`, `cancelled`, `queue_timeout`, `request_not_found`, `queue_backlog_full` and `result_expired`. Closed describes the set as documented today, not a bound that holds forever: the set is expected to grow, which is why this is deliberately a plain string and not an `enum`, so a client must treat an unrecognised value as `internal_error` rather than switch exhaustively over the list above and break on the next addition.","example":"invalid_input","x-comfy-error-types":[{"value":"invalid_input","tier":"request","meaning":"The request was rejected before it reached the model - a malformed body, a malformed or expired pagination cursor, an input the model's own schema does not accept, or an `Idempotency-Key` that cannot serve this request (already used for a different request - the method, the path and query, or the body differ - or already consumed by a call whose response cannot be replayed). Sent with `409` in the key cases and with `400`/`422` in the others; the status says which, and the key cases are the ones answered by using a new key rather than by editing the request."},{"value":"content_policy_violation","tier":"request","meaning":"The provider refused the request on content-policy grounds. The refusal is deterministic: re-sending the same input will be refused again."},{"value":"provider_error","tier":"request","meaning":"The partner provider reported a failure of its own, or returned a response Router could not interpret as a result."},{"value":"provider_timeout","tier":"request","meaning":"The partner provider did not answer within its deadline. This bucket is the provider timing out and never Router's own server deadline, which is reported as `deadline_exceeded` - the two share `504` and are separated because they name different causes: this one says the partner failed, that one says Comfy stopped holding the connection."},{"value":"insufficient_credits","tier":"request","meaning":"The calling workspace does not have enough credits to run the model."},{"value":"model_not_found","tier":"request","meaning":"The `{provider}/{model}` ID names no model Router can run: an unknown provider, an unenrolled id, one the catalog has retired past its sunset date, or a catalogued ID the provider does not currently serve for Comfy. A retired id's `detail` reads `\u003cid> was retired on \u003cYYYY-MM-DD>; use \u003csuccessor>` when a successor is recorded and not withheld, or just `\u003cid> was retired on \u003cYYYY-MM-DD>` when it is not, instead of a suggestion list; the successor named for a retired id is always a live catalog ID you can re-point the call at, and a successor is dropped from `detail` either because the table records none, or because the recorded one is currently withheld behind an embargo the caller cannot see. A model whose sunset is still in the future is not withdrawn: it stays listed by `GET /v2/models` and stays runnable until the date, so a listing taken today remains the answer to what runs today. Otherwise `detail` carries up to three suggestions drawn from the models the caller is entitled to see, except for a catalogued ID the provider does not currently serve for Comfy, whose response carries no suggestions."},{"value":"unauthorized","tier":"transport","meaning":"The request carried no usable credential."},{"value":"forbidden","tier":"transport","meaning":"The credential is valid but is not entitled to this model or this operation."},{"value":"concurrency_limit_exceeded","tier":"transport","meaning":"The workspace already has as many calls in flight as it is allowed; retry once one of them finishes. It carries one further condition on the run route, on a `409` rather than the `429` above: another call is already in flight for the `Idempotency-Key` this request presented. Re-send the same key after `Retry-After` seconds to collect that call's result."},{"value":"client_disconnected","tier":"transport","meaning":"The caller closed the connection before Router could return a result. It is logged rather than delivered - there is no socket left to write it to - and it is an attribution, not a billing outcome: a provider generation that completed is billed regardless of whether the caller received the response."},{"value":"internal_error","tier":"transport","meaning":"Router itself failed. It is also the value a client should treat any unrecognized bucket as, so a later addition to the set does not break a client generated before it."},{"value":"deadline_exceeded","tier":"transport","meaning":"Comfy stopped holding the connection at its own configured bound before an answer arrived. It shares `504` with `provider_timeout` and the pair says which side ran out of time; this one is Comfy's own bound, so nothing about the request was rejected and the same request may be retried. It says nothing about the charge: a provider generation that completed is billed regardless of whether the caller received the response. Retry it with the same `Idempotency-Key`: when the provider had already accepted the generation, the retry collects that generation rather than dispatching another, and a `Retry-After` on the `504` says when to ask."},{"value":"not_enabled","tier":"transport","meaning":"Comfy Router is not switched on for this caller yet. Nothing about the request is wrong and the model exists, which is why this is not `model_not_found`; it shares `403` with `forbidden` and is not the same thing, because `forbidden` is an entitlement decision about the caller while this is a state of the rollout. It is terminal: do not retry, and do not treat it as an outage. The one exception to \"about the caller\" is the queued submit, which also answers `not_enabled` for a model whose partner answers a generation directly as bytes: that model cannot yet be queued, so it is the model and not the caller that is refused, nothing is queued or charged, and the synchronous route `POST /v2/models/{provider}/{model}` runs it instead."},{"value":"service_unavailable","tier":"transport","meaning":"A service Comfy Router depends on is temporarily unavailable and the caller did nothing wrong. Retry it with backoff: it is the one bucket here whose condition clears on its own, without the caller changing the request and without a concurrency slot freeing, which is what distinguishes it from the other retryable answers (`concurrency_limit_exceeded`, `deadline_exceeded`). It is separate from `internal_error` - which is a `500` and means Router itself failed - so a client can tell \"come back shortly\" from \"this call is not going to work\"."},{"value":"rate_limited","tier":"transport","meaning":"The caller has spent an allowance measured over a window and must wait for that window to roll. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one clears the moment one of the caller's own in-flight calls finishes, so retrying in seconds is right, whereas nothing the caller does drains this one early. `detail` names the window, and `Retry-After` the seconds to wait."},{"value":"cancelled","tier":"transport","meaning":"A queued request was withdrawn — through the cancel route, or by an operator — before it produced a result; it is terminal, and it is not by itself a statement about the charge. Cancelling stops Comfy waiting and it does not always stop the partner working, so a request cancelled while it was still `IN_QUEUE` was never dispatched and cannot be charged, whereas one cancelled after it was admitted may still be charged — a partner generation that completes is charged whether or not anyone collected it, which is why the cancel route calls the ask a request rather than a guarantee. It is not `client_disconnected`: that one says nobody is listening any more while a generation may still be running and billable, whereas this says the request itself was withdrawn. A caller polling the status read sees `200` with this in the body, because the request ended and so the read succeeded. Collecting the result of such a request answers `409` in Router's error envelope: that read also worked, and what it found is a request whose terminal state - the caller's own decision - leaves nothing to return. It is deliberately not `410`, which on that route means the result aged out of its retention window, and not a `5xx`, which would report the caller's own cancellation as a Router fault and read to a polling client as worth retrying."},{"value":"queue_timeout","tier":"transport","meaning":"A queued request waited past its queue timeout without ever being admitted. Terminal, unbilled, and it never took a concurrency slot — the job never reached a provider. Deliberately not `deadline_exceeded`, which is the synchronous route's connection bound: a `deadline_exceeded` generation may be running and billable, whereas this one provably never started. It is returned under `504`, which it shares with `provider_timeout` and `deadline_exceeded` because a status can only say that a clock ran out; which clock - the partner's, Comfy's connection bound, or the queue's admission bound - is what `error_type` carries. The request itself is terminal, so submit a new one rather than re-reading this one."},{"value":"request_not_found","tier":"transport","meaning":"The `request_id` names no request of the caller's under this model. It is the second of the two conditions the queued reads' `404` covers; the first is the `{provider}/{model}` ID resolving to no partner model, which is `model_not_found` and carries fuzzy model suggestions. It also covers the right-id / wrong-model URL the path shape refuses, and it is deliberately indistinguishable from a request in another workspace, so a probe with a guessed id learns nothing. A request that has merely aged out of its retention window is `410`, not this."},{"value":"queue_backlog_full","tier":"transport","meaning":"The caller already has too many queued requests waiting to run, so this submit was refused. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one is the synchronous route's answer for too many calls in flight at once, whereas the queue accepts a submit at that limit and parks it, and this bucket is the separate bound on how many a caller may leave waiting so that parking cannot mean enqueuing without end. It clears as the caller's own queued requests finish, so retry once some of them complete."},{"value":"result_expired","tier":"transport","meaning":"The request existed and finished, but its result is no longer stored, so the collect has nothing to return. It is returned under `410` and is permanent for that id: re-reading will answer the same way, so do not retry it. It is deliberately not `request_not_found`, which would tell you the id was never valid and invite re-submitting a generation that has already run. Running the model again means submitting a new request. For a workspace under a zero-data-retention agreement, a queued request's result is not kept after it is delivered: it is removed shortly after the first collect, and an uncollected result is removed after a grace period once the request finishes, so store the result on your side when you first collect it."}],"isRequired":true,"uniqueKey":"error_type","typeLabel":"string"},"upstream_detail":{"type":"string","description":"A bounded, sanitized reason the model provider gave for rejecting the request, present only when `error_type` is `invalid_input` and `X-Comfy-Upstream-Status` is a provider `4xx` or `2xx` - i.e. the provider refused the request as malformed and said why. Usually that status is a `4xx`; it is a `2xx` for a provider that reports a rejected generation inside a success envelope (a BytePlus failed-task poll is HTTP `200` with the reason in its body). Absent on every other failure, including provider `5xx`, transport failures, content-policy refusals and any refusal Router raised about itself. It mirrors the `X-Comfy-Upstream-Detail` header. The queued collect (`GET /v2/models/{provider}/{model}/requests/{request_id}`) of a run that ended `invalid_input` carries the same reason in this field on its first collect, which then deletes it from the request: a later collect normally answers the same `error_type` and `detail` without it, so read it from the first collect. That read sends neither header.","uniqueKey":"upstream_detail","typeLabel":"string"},"refusal_subject":{"type":"string","description":"Which input or output a content-policy refusal was about, as a Router-level closed vocabulary: `input`, `output`, `input_text`, `input_image`, `input_video`, `input_audio`, `output_text`, `output_image`, `output_video`, `output_audio`. The bare `input` / `output` values name the side when the provider did not name a modality. Present only when `error_type` is `content_policy_violation` and the provider named the refused subject with a machine-readable code; absent otherwise. Never provider text. Named today for BytePlus, Runway, BFL, Gemini, Veo, Vertex, xAI and Wan refusals, and for the MiniMax Router's Nano Banana legs' Gemini refusals; a provider whose refusal does not say which side it was about leaves it absent. It mirrors the `X-Comfy-Refusal-Subject` header.","uniqueKey":"refusal_subject","typeLabel":"string"},"denied_reason":{"type":"string","description":"Billing's policy branch for a partner-node spend denial - `no_funds` when the billed wallet is empty, `unspecified` when the call was refused without a funds reading saying so. Present only when `error_type` is `insufficient_credits` and the deployment reports wallet facts; absent otherwise. A machine-readable code, not display text: show `detail`, branch on this.","uniqueKey":"denied_reason","typeLabel":"string"},"billed_workspace":{"type":"object","description":"The wallet a refused call would have charged, on `RouterErrorResponse.billed_workspace`. Answers \"which wallet is empty?\" - it is not necessarily the workspace the caller's UI currently has selected, and a team member's personal credential names their personal wallet here even while a funded team is open. Present only alongside `denied_reason`.","properties":{"type":{"type":"string","description":"`personal` or `team`. An open string: treat any other value as personal.","isRequired":true,"uniqueKey":"billed_workspace.type","typeLabel":"string"},"id":{"type":"string","description":"The Cloud workspace id. Absent on the legacy rail.","uniqueKey":"billed_workspace.id","typeLabel":"string"},"rail":{"type":"string","description":"`legacy` when the call bills the caller's pre-workspace Metronome customer rather than a Cloud workspace; that customer has no workspace id. Absent for a Cloud workspace.","uniqueKey":"billed_workspace.rail","typeLabel":"string"}},"required":["type"],"uniqueKey":"billed_workspace","typeLabel":"object"},"credential_kind":{"type":"string","description":"How the refused request authenticated - `personal_api_key`, `team_api_key`, `personal_session` or `team_session`. Answers \"which credential do I switch to bill the other wallet?\". It is not a permission and grants nothing. Present only alongside `denied_reason`.","uniqueKey":"credential_kind","typeLabel":"string"}},"required":["detail","error_type"],"uniqueKey":"","typeLabel":"object"},"examples":{"Example":{"title":"Example","value":{"detail":"\u003cstring>","error_type":"invalid_input","upstream_detail":"\u003cstring>","refusal_subject":"\u003cstring>","denied_reason":"\u003cstring>","billed_workspace":{"type":"\u003cstring>","id":"\u003cstring>","rail":"\u003cstring>"},"credential_kind":"\u003cstring>"}}},"exampleType":"generated"}},"exampleType":"generated"},"a2bbb60a-4833-42cc-a11e-be4f6a942588":{"description":"A Router request-level failure - the request never reached the model, or failed for a reason the model itself did not report. The body is `RouterErrorResponse` and the bucket is repeated on `X-Comfy-Error-Type`.","headers":{"X-Comfy-Error-Type":{"description":"Coarse, machine-readable bucket for the failure, set by Router on every error response. It carries the same value as `RouterErrorResponse.error_type`, and on the `422` it is the only machine-readable bucket, because that body is the FastAPI `detail[]` shape and has no `error_type` field of its own. A client can therefore branch on this header alone, before deciding which of the two Router error bodies it received.","required":true,"schema":{"type":"string","description":"Coarse, machine-readable bucket for a Router failure, mirrored on the `X-Comfy-Error-Type` response header so a caller can branch without parsing the body. The set is closed at twenty values: the six request-level buckets `invalid_input`, `content_policy_violation`, `provider_error`, `provider_timeout`, `insufficient_credits` and `model_not_found`, plus the transport-level `unauthorized`, `forbidden`, `concurrency_limit_exceeded`, `client_disconnected`, `internal_error`, `deadline_exceeded`, `not_enabled`, `service_unavailable`, `rate_limited`, `cancelled`, `queue_timeout`, `request_not_found`, `queue_backlog_full` and `result_expired`. Closed describes the set as documented today, not a bound that holds forever: the set is expected to grow, which is why this is deliberately a plain string and not an `enum`, so a client must treat an unrecognised value as `internal_error` rather than switch exhaustively over the list above and break on the next addition.","example":"invalid_input","x-comfy-error-types":[{"value":"invalid_input","tier":"request","meaning":"The request was rejected before it reached the model - a malformed body, a malformed or expired pagination cursor, an input the model's own schema does not accept, or an `Idempotency-Key` that cannot serve this request (already used for a different request - the method, the path and query, or the body differ - or already consumed by a call whose response cannot be replayed). Sent with `409` in the key cases and with `400`/`422` in the others; the status says which, and the key cases are the ones answered by using a new key rather than by editing the request."},{"value":"content_policy_violation","tier":"request","meaning":"The provider refused the request on content-policy grounds. The refusal is deterministic: re-sending the same input will be refused again."},{"value":"provider_error","tier":"request","meaning":"The partner provider reported a failure of its own, or returned a response Router could not interpret as a result."},{"value":"provider_timeout","tier":"request","meaning":"The partner provider did not answer within its deadline. This bucket is the provider timing out and never Router's own server deadline, which is reported as `deadline_exceeded` - the two share `504` and are separated because they name different causes: this one says the partner failed, that one says Comfy stopped holding the connection."},{"value":"insufficient_credits","tier":"request","meaning":"The calling workspace does not have enough credits to run the model."},{"value":"model_not_found","tier":"request","meaning":"The `{provider}/{model}` ID names no model Router can run: an unknown provider, an unenrolled id, one the catalog has retired past its sunset date, or a catalogued ID the provider does not currently serve for Comfy. A retired id's `detail` reads `\u003cid> was retired on \u003cYYYY-MM-DD>; use \u003csuccessor>` when a successor is recorded and not withheld, or just `\u003cid> was retired on \u003cYYYY-MM-DD>` when it is not, instead of a suggestion list; the successor named for a retired id is always a live catalog ID you can re-point the call at, and a successor is dropped from `detail` either because the table records none, or because the recorded one is currently withheld behind an embargo the caller cannot see. A model whose sunset is still in the future is not withdrawn: it stays listed by `GET /v2/models` and stays runnable until the date, so a listing taken today remains the answer to what runs today. Otherwise `detail` carries up to three suggestions drawn from the models the caller is entitled to see, except for a catalogued ID the provider does not currently serve for Comfy, whose response carries no suggestions."},{"value":"unauthorized","tier":"transport","meaning":"The request carried no usable credential."},{"value":"forbidden","tier":"transport","meaning":"The credential is valid but is not entitled to this model or this operation."},{"value":"concurrency_limit_exceeded","tier":"transport","meaning":"The workspace already has as many calls in flight as it is allowed; retry once one of them finishes. It carries one further condition on the run route, on a `409` rather than the `429` above: another call is already in flight for the `Idempotency-Key` this request presented. Re-send the same key after `Retry-After` seconds to collect that call's result."},{"value":"client_disconnected","tier":"transport","meaning":"The caller closed the connection before Router could return a result. It is logged rather than delivered - there is no socket left to write it to - and it is an attribution, not a billing outcome: a provider generation that completed is billed regardless of whether the caller received the response."},{"value":"internal_error","tier":"transport","meaning":"Router itself failed. It is also the value a client should treat any unrecognized bucket as, so a later addition to the set does not break a client generated before it."},{"value":"deadline_exceeded","tier":"transport","meaning":"Comfy stopped holding the connection at its own configured bound before an answer arrived. It shares `504` with `provider_timeout` and the pair says which side ran out of time; this one is Comfy's own bound, so nothing about the request was rejected and the same request may be retried. It says nothing about the charge: a provider generation that completed is billed regardless of whether the caller received the response. Retry it with the same `Idempotency-Key`: when the provider had already accepted the generation, the retry collects that generation rather than dispatching another, and a `Retry-After` on the `504` says when to ask."},{"value":"not_enabled","tier":"transport","meaning":"Comfy Router is not switched on for this caller yet. Nothing about the request is wrong and the model exists, which is why this is not `model_not_found`; it shares `403` with `forbidden` and is not the same thing, because `forbidden` is an entitlement decision about the caller while this is a state of the rollout. It is terminal: do not retry, and do not treat it as an outage. The one exception to \"about the caller\" is the queued submit, which also answers `not_enabled` for a model whose partner answers a generation directly as bytes: that model cannot yet be queued, so it is the model and not the caller that is refused, nothing is queued or charged, and the synchronous route `POST /v2/models/{provider}/{model}` runs it instead."},{"value":"service_unavailable","tier":"transport","meaning":"A service Comfy Router depends on is temporarily unavailable and the caller did nothing wrong. Retry it with backoff: it is the one bucket here whose condition clears on its own, without the caller changing the request and without a concurrency slot freeing, which is what distinguishes it from the other retryable answers (`concurrency_limit_exceeded`, `deadline_exceeded`). It is separate from `internal_error` - which is a `500` and means Router itself failed - so a client can tell \"come back shortly\" from \"this call is not going to work\"."},{"value":"rate_limited","tier":"transport","meaning":"The caller has spent an allowance measured over a window and must wait for that window to roll. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one clears the moment one of the caller's own in-flight calls finishes, so retrying in seconds is right, whereas nothing the caller does drains this one early. `detail` names the window, and `Retry-After` the seconds to wait."},{"value":"cancelled","tier":"transport","meaning":"A queued request was withdrawn — through the cancel route, or by an operator — before it produced a result; it is terminal, and it is not by itself a statement about the charge. Cancelling stops Comfy waiting and it does not always stop the partner working, so a request cancelled while it was still `IN_QUEUE` was never dispatched and cannot be charged, whereas one cancelled after it was admitted may still be charged — a partner generation that completes is charged whether or not anyone collected it, which is why the cancel route calls the ask a request rather than a guarantee. It is not `client_disconnected`: that one says nobody is listening any more while a generation may still be running and billable, whereas this says the request itself was withdrawn. A caller polling the status read sees `200` with this in the body, because the request ended and so the read succeeded. Collecting the result of such a request answers `409` in Router's error envelope: that read also worked, and what it found is a request whose terminal state - the caller's own decision - leaves nothing to return. It is deliberately not `410`, which on that route means the result aged out of its retention window, and not a `5xx`, which would report the caller's own cancellation as a Router fault and read to a polling client as worth retrying."},{"value":"queue_timeout","tier":"transport","meaning":"A queued request waited past its queue timeout without ever being admitted. Terminal, unbilled, and it never took a concurrency slot — the job never reached a provider. Deliberately not `deadline_exceeded`, which is the synchronous route's connection bound: a `deadline_exceeded` generation may be running and billable, whereas this one provably never started. It is returned under `504`, which it shares with `provider_timeout` and `deadline_exceeded` because a status can only say that a clock ran out; which clock - the partner's, Comfy's connection bound, or the queue's admission bound - is what `error_type` carries. The request itself is terminal, so submit a new one rather than re-reading this one."},{"value":"request_not_found","tier":"transport","meaning":"The `request_id` names no request of the caller's under this model. It is the second of the two conditions the queued reads' `404` covers; the first is the `{provider}/{model}` ID resolving to no partner model, which is `model_not_found` and carries fuzzy model suggestions. It also covers the right-id / wrong-model URL the path shape refuses, and it is deliberately indistinguishable from a request in another workspace, so a probe with a guessed id learns nothing. A request that has merely aged out of its retention window is `410`, not this."},{"value":"queue_backlog_full","tier":"transport","meaning":"The caller already has too many queued requests waiting to run, so this submit was refused. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one is the synchronous route's answer for too many calls in flight at once, whereas the queue accepts a submit at that limit and parks it, and this bucket is the separate bound on how many a caller may leave waiting so that parking cannot mean enqueuing without end. It clears as the caller's own queued requests finish, so retry once some of them complete."},{"value":"result_expired","tier":"transport","meaning":"The request existed and finished, but its result is no longer stored, so the collect has nothing to return. It is returned under `410` and is permanent for that id: re-reading will answer the same way, so do not retry it. It is deliberately not `request_not_found`, which would tell you the id was never valid and invite re-submitting a generation that has already run. Running the model again means submitting a new request. For a workspace under a zero-data-retention agreement, a queued request's result is not kept after it is delivered: it is removed shortly after the first collect, and an uncollected result is removed after a grace period once the request finishes, so store the result on your side when you first collect it."}]}},"X-Comfy-Request-Id":{"description":"Server-generated identifier for this call, present on every Router response - success, 4xx and 5xx alike, because an error response is exactly when a user needs an id to quote in a support request. The same value is written into the call's usage/audit event, which is what lets a complaint about a charge be joined to the charge itself instead of searched for by timestamp.","required":true,"schema":{"type":"string","format":"uuid","example":"6f1a1a6e-6a53-4a5f-9d3a-2b3b0a1f9c21"}}},"content":{"application/json":{"schema":{"type":"object","description":"Router's request-level error body: what is returned when the request never reached the model, or failed for a reason the model itself did not report - auth, quota, an unknown model ID, or provider transport. A model-level validation failure has its own shape, `RouterValidationErrorResponse`, because flattening a FastAPI `detail[]` array into this `detail` string would destroy the per-field granularity an SDK branches on.","properties":{"detail":{"type":"string","description":"Human-readable description of the failure, safe to surface to an end user. Not machine-parsed - branch on `error_type` instead. For `model_not_found` on a retired model, `detail` reads `\u003cid> was retired on \u003cYYYY-MM-DD>; use \u003csuccessor>` when a successor is recorded and visible to the caller, or just `\u003cid> was retired on \u003cYYYY-MM-DD>` when it is not, instead of the usual suggestion list. A successor named here is always a live catalog ID the call can be re-pointed at. A model whose sunset date is still in the future is not retired: it stays listed by `GET /v2/models` and stays runnable until that date. Like the rest of `detail`, this retirement wording is for display and carries no compatibility guarantee.","isRequired":true,"uniqueKey":"detail","typeLabel":"string"},"error_type":{"type":"string","description":"Coarse, machine-readable bucket for a Router failure, mirrored on the `X-Comfy-Error-Type` response header so a caller can branch without parsing the body. The set is closed at twenty values: the six request-level buckets `invalid_input`, `content_policy_violation`, `provider_error`, `provider_timeout`, `insufficient_credits` and `model_not_found`, plus the transport-level `unauthorized`, `forbidden`, `concurrency_limit_exceeded`, `client_disconnected`, `internal_error`, `deadline_exceeded`, `not_enabled`, `service_unavailable`, `rate_limited`, `cancelled`, `queue_timeout`, `request_not_found`, `queue_backlog_full` and `result_expired`. Closed describes the set as documented today, not a bound that holds forever: the set is expected to grow, which is why this is deliberately a plain string and not an `enum`, so a client must treat an unrecognised value as `internal_error` rather than switch exhaustively over the list above and break on the next addition.","example":"invalid_input","x-comfy-error-types":[{"value":"invalid_input","tier":"request","meaning":"The request was rejected before it reached the model - a malformed body, a malformed or expired pagination cursor, an input the model's own schema does not accept, or an `Idempotency-Key` that cannot serve this request (already used for a different request - the method, the path and query, or the body differ - or already consumed by a call whose response cannot be replayed). Sent with `409` in the key cases and with `400`/`422` in the others; the status says which, and the key cases are the ones answered by using a new key rather than by editing the request."},{"value":"content_policy_violation","tier":"request","meaning":"The provider refused the request on content-policy grounds. The refusal is deterministic: re-sending the same input will be refused again."},{"value":"provider_error","tier":"request","meaning":"The partner provider reported a failure of its own, or returned a response Router could not interpret as a result."},{"value":"provider_timeout","tier":"request","meaning":"The partner provider did not answer within its deadline. This bucket is the provider timing out and never Router's own server deadline, which is reported as `deadline_exceeded` - the two share `504` and are separated because they name different causes: this one says the partner failed, that one says Comfy stopped holding the connection."},{"value":"insufficient_credits","tier":"request","meaning":"The calling workspace does not have enough credits to run the model."},{"value":"model_not_found","tier":"request","meaning":"The `{provider}/{model}` ID names no model Router can run: an unknown provider, an unenrolled id, one the catalog has retired past its sunset date, or a catalogued ID the provider does not currently serve for Comfy. A retired id's `detail` reads `\u003cid> was retired on \u003cYYYY-MM-DD>; use \u003csuccessor>` when a successor is recorded and not withheld, or just `\u003cid> was retired on \u003cYYYY-MM-DD>` when it is not, instead of a suggestion list; the successor named for a retired id is always a live catalog ID you can re-point the call at, and a successor is dropped from `detail` either because the table records none, or because the recorded one is currently withheld behind an embargo the caller cannot see. A model whose sunset is still in the future is not withdrawn: it stays listed by `GET /v2/models` and stays runnable until the date, so a listing taken today remains the answer to what runs today. Otherwise `detail` carries up to three suggestions drawn from the models the caller is entitled to see, except for a catalogued ID the provider does not currently serve for Comfy, whose response carries no suggestions."},{"value":"unauthorized","tier":"transport","meaning":"The request carried no usable credential."},{"value":"forbidden","tier":"transport","meaning":"The credential is valid but is not entitled to this model or this operation."},{"value":"concurrency_limit_exceeded","tier":"transport","meaning":"The workspace already has as many calls in flight as it is allowed; retry once one of them finishes. It carries one further condition on the run route, on a `409` rather than the `429` above: another call is already in flight for the `Idempotency-Key` this request presented. Re-send the same key after `Retry-After` seconds to collect that call's result."},{"value":"client_disconnected","tier":"transport","meaning":"The caller closed the connection before Router could return a result. It is logged rather than delivered - there is no socket left to write it to - and it is an attribution, not a billing outcome: a provider generation that completed is billed regardless of whether the caller received the response."},{"value":"internal_error","tier":"transport","meaning":"Router itself failed. It is also the value a client should treat any unrecognized bucket as, so a later addition to the set does not break a client generated before it."},{"value":"deadline_exceeded","tier":"transport","meaning":"Comfy stopped holding the connection at its own configured bound before an answer arrived. It shares `504` with `provider_timeout` and the pair says which side ran out of time; this one is Comfy's own bound, so nothing about the request was rejected and the same request may be retried. It says nothing about the charge: a provider generation that completed is billed regardless of whether the caller received the response. Retry it with the same `Idempotency-Key`: when the provider had already accepted the generation, the retry collects that generation rather than dispatching another, and a `Retry-After` on the `504` says when to ask."},{"value":"not_enabled","tier":"transport","meaning":"Comfy Router is not switched on for this caller yet. Nothing about the request is wrong and the model exists, which is why this is not `model_not_found`; it shares `403` with `forbidden` and is not the same thing, because `forbidden` is an entitlement decision about the caller while this is a state of the rollout. It is terminal: do not retry, and do not treat it as an outage. The one exception to \"about the caller\" is the queued submit, which also answers `not_enabled` for a model whose partner answers a generation directly as bytes: that model cannot yet be queued, so it is the model and not the caller that is refused, nothing is queued or charged, and the synchronous route `POST /v2/models/{provider}/{model}` runs it instead."},{"value":"service_unavailable","tier":"transport","meaning":"A service Comfy Router depends on is temporarily unavailable and the caller did nothing wrong. Retry it with backoff: it is the one bucket here whose condition clears on its own, without the caller changing the request and without a concurrency slot freeing, which is what distinguishes it from the other retryable answers (`concurrency_limit_exceeded`, `deadline_exceeded`). It is separate from `internal_error` - which is a `500` and means Router itself failed - so a client can tell \"come back shortly\" from \"this call is not going to work\"."},{"value":"rate_limited","tier":"transport","meaning":"The caller has spent an allowance measured over a window and must wait for that window to roll. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one clears the moment one of the caller's own in-flight calls finishes, so retrying in seconds is right, whereas nothing the caller does drains this one early. `detail` names the window, and `Retry-After` the seconds to wait."},{"value":"cancelled","tier":"transport","meaning":"A queued request was withdrawn — through the cancel route, or by an operator — before it produced a result; it is terminal, and it is not by itself a statement about the charge. Cancelling stops Comfy waiting and it does not always stop the partner working, so a request cancelled while it was still `IN_QUEUE` was never dispatched and cannot be charged, whereas one cancelled after it was admitted may still be charged — a partner generation that completes is charged whether or not anyone collected it, which is why the cancel route calls the ask a request rather than a guarantee. It is not `client_disconnected`: that one says nobody is listening any more while a generation may still be running and billable, whereas this says the request itself was withdrawn. A caller polling the status read sees `200` with this in the body, because the request ended and so the read succeeded. Collecting the result of such a request answers `409` in Router's error envelope: that read also worked, and what it found is a request whose terminal state - the caller's own decision - leaves nothing to return. It is deliberately not `410`, which on that route means the result aged out of its retention window, and not a `5xx`, which would report the caller's own cancellation as a Router fault and read to a polling client as worth retrying."},{"value":"queue_timeout","tier":"transport","meaning":"A queued request waited past its queue timeout without ever being admitted. Terminal, unbilled, and it never took a concurrency slot — the job never reached a provider. Deliberately not `deadline_exceeded`, which is the synchronous route's connection bound: a `deadline_exceeded` generation may be running and billable, whereas this one provably never started. It is returned under `504`, which it shares with `provider_timeout` and `deadline_exceeded` because a status can only say that a clock ran out; which clock - the partner's, Comfy's connection bound, or the queue's admission bound - is what `error_type` carries. The request itself is terminal, so submit a new one rather than re-reading this one."},{"value":"request_not_found","tier":"transport","meaning":"The `request_id` names no request of the caller's under this model. It is the second of the two conditions the queued reads' `404` covers; the first is the `{provider}/{model}` ID resolving to no partner model, which is `model_not_found` and carries fuzzy model suggestions. It also covers the right-id / wrong-model URL the path shape refuses, and it is deliberately indistinguishable from a request in another workspace, so a probe with a guessed id learns nothing. A request that has merely aged out of its retention window is `410`, not this."},{"value":"queue_backlog_full","tier":"transport","meaning":"The caller already has too many queued requests waiting to run, so this submit was refused. It shares `429` with `concurrency_limit_exceeded` and is not the same thing: that one is the synchronous route's answer for too many calls in flight at once, whereas the queue accepts a submit at that limit and parks it, and this bucket is the separate bound on how many a caller may leave waiting so that parking cannot mean enqueuing without end. It clears as the caller's own queued requests finish, so retry once some of them complete."},{"value":"result_expired","tier":"transport","meaning":"The request existed and finished, but its result is no longer stored, so the collect has nothing to return. It is returned under `410` and is permanent for that id: re-reading will answer the same way, so do not retry it. It is deliberately not `request_not_found`, which would tell you the id was never valid and invite re-submitting a generation that has already run. Running the model again means submitting a new request. For a workspace under a zero-data-retention agreement, a queued request's result is not kept after it is delivered: it is removed shortly after the first collect, and an uncollected result is removed after a grace period once the request finishes, so store the result on your side when you first collect it."}],"isRequired":true,"uniqueKey":"error_type","typeLabel":"string"},"upstream_detail":{"type":"string","description":"A bounded, sanitized reason the model provider gave for rejecting the request, present only when `error_type` is `invalid_input` and `X-Comfy-Upstream-Status` is a provider `4xx` or `2xx` - i.e. the provider refused the request as malformed and said why. Usually that status is a `4xx`; it is a `2xx` for a provider that reports a rejected generation inside a success envelope (a BytePlus failed-task poll is HTTP `200` with the reason in its body). Absent on every other failure, including provider `5xx`, transport failures, content-policy refusals and any refusal Router raised about itself. It mirrors the `X-Comfy-Upstream-Detail` header. The queued collect (`GET /v2/models/{provider}/{model}/requests/{request_id}`) of a run that ended `invalid_input` carries the same reason in this field on its first collect, which then deletes it from the request: a later collect normally answers the same `error_type` and `detail` without it, so read it from the first collect. That read sends neither header.","uniqueKey":"upstream_detail","typeLabel":"string"},"refusal_subject":{"type":"string","description":"Which input or output a content-policy refusal was about, as a Router-level closed vocabulary: `input`, `output`, `input_text`, `input_image`, `input_video`, `input_audio`, `output_text`, `output_image`, `output_video`, `output_audio`. The bare `input` / `output` values name the side when the provider did not name a modality. Present only when `error_type` is `content_policy_violation` and the provider named the refused subject with a machine-readable code; absent otherwise. Never provider text. Named today for BytePlus, Runway, BFL, Gemini, Veo, Vertex, xAI and Wan refusals, and for the MiniMax Router's Nano Banana legs' Gemini refusals; a provider whose refusal does not say which side it was about leaves it absent. It mirrors the `X-Comfy-Refusal-Subject` header.","uniqueKey":"refusal_subject","typeLabel":"string"},"denied_reason":{"type":"string","description":"Billing's policy branch for a partner-node spend denial - `no_funds` when the billed wallet is empty, `unspecified` when the call was refused without a funds reading saying so. Present only when `error_type` is `insufficient_credits` and the deployment reports wallet facts; absent otherwise. A machine-readable code, not display text: show `detail`, branch on this.","uniqueKey":"denied_reason","typeLabel":"string"},"billed_workspace":{"type":"object","description":"The wallet a refused call would have charged, on `RouterErrorResponse.billed_workspace`. Answers \"which wallet is empty?\" - it is not necessarily the workspace the caller's UI currently has selected, and a team member's personal credential names their personal wallet here even while a funded team is open. Present only alongside `denied_reason`.","properties":{"type":{"type":"string","description":"`personal` or `team`. An open string: treat any other value as personal.","isRequired":true,"uniqueKey":"billed_workspace.type","typeLabel":"string"},"id":{"type":"string","description":"The Cloud workspace id. Absent on the legacy rail.","uniqueKey":"billed_workspace.id","typeLabel":"string"},"rail":{"type":"string","description":"`legacy` when the call bills the caller's pre-workspace Metronome customer rather than a Cloud workspace; that customer has no workspace id. Absent for a Cloud workspace.","uniqueKey":"billed_workspace.rail","typeLabel":"string"}},"required":["type"],"uniqueKey":"billed_workspace","typeLabel":"object"},"credential_kind":{"type":"string","description":"How the refused request authenticated - `personal_api_key`, `team_api_key`, `personal_session` or `team_session`. Answers \"which credential do I switch to bill the other wallet?\". It is not a permission and grants nothing. Present only alongside `denied_reason`.","uniqueKey":"credential_kind","typeLabel":"string"}},"required":["detail","error_type"],"uniqueKey":"","typeLabel":"object"},"examples":{"Example":{"title":"Example","value":{"detail":"\u003cstring>","error_type":"invalid_input","upstream_detail":"\u003cstring>","refusal_subject":"\u003cstring>","denied_reason":"\u003cstring>","billed_workspace":{"type":"\u003cstring>","id":"\u003cstring>","rail":"\u003cstring>"},"credential_kind":"\u003cstring>"}}},"exampleType":"generated"}},"exampleType":"generated"}},"security":{"e7e3f8e5-bea5-4030-b1db-8f46dca76241":{"BearerAuth":{"type":"http","scheme":"bearer","bearerFormat":"JWT","description":"Bearer token authentication. Normally a Firebase or Cloud JWT. A 'comfyui-' prefixed API key is also accepted in this header: the prefix classifies the value as an API key and it is validated exactly as if it had been sent in X-API-Key.","scopes":[],"schemeName":"BearerAuth"}},"3ad52369-c0be-4ebc-8842-96afd522dd68":{"ApiKeyAuth":{"type":"apiKey","in":"header","name":"X-API-Key","description":"API key authentication. Send the key in the X-API-Key header; keys are prefixed with 'comfyui-' and are generated from user account settings. The same 'comfyui-' key is also accepted in Authorization: Bearer (see BearerAuth), and when both headers carry a key, X-API-Key wins.","scopes":[],"schemeName":"ApiKeyAuth"}}},"servers":{"bf9722b0-5952-4eeb-9e69-c43ede83ce5a":{"url":"https://api.comfy.org"}},"processedSecurityOptions":[{"schemes":[{"scheme":{"type":"http","scheme":"bearer","bearerFormat":"JWT","description":"Bearer token authentication. Normally a Firebase or Cloud JWT. A 'comfyui-' prefixed API key is also accepted in this header: the prefix classifies the value as an API key and it is validated exactly as if it had been sent in X-API-Key.","scopes":[],"schemeName":"BearerAuth"},"schema":{"title":"Bearer Auth","type":"string","isRequired":true,"description":"Bearer token authentication. Normally a Firebase or Cloud JWT. A 'comfyui-' prefixed API key is also accepted in this header: the prefix classifies the value as an API key and it is validated exactly as if it had been sent in X-API-Key.","format":"bearer","typeLabel":"string\u003cbearer>","deprecated":false},"schemeName":"BearerAuth","name":"Authorization"}],"title":"BearerAuth"},{"schemes":[{"scheme":{"type":"apiKey","in":"header","name":"X-API-Key","description":"API key authentication. Send the key in the X-API-Key header; keys are prefixed with 'comfyui-' and are generated from user account settings. The same 'comfyui-' key is also accepted in Authorization: Bearer (see BearerAuth), and when both headers carry a key, X-API-Key wins.","scopes":[],"schemeName":"ApiKeyAuth"},"schema":{"title":"X-API-Key","type":"string","description":"API key authentication. Send the key in the X-API-Key header; keys are prefixed with 'comfyui-' and are generated from user account settings. The same 'comfyui-' key is also accepted in Authorization: Bearer (see BearerAuth), and when both headers carry a key, X-API-Key wins.","isRequired":true,"typeLabel":"string","deprecated":false},"schemeName":"ApiKeyAuth","name":"X-API-Key"}],"title":"ApiKeyAuth"}]},"originalFileLocation":"router-openapi.yaml"},"subdomain":"dripart","siteName":"ComfyUI","title":"List the models Comfy Router can run. - ComfyUI","navMaps":{"firstHrefInVersion":{},"firstHrefInLanguage":{"en":{"href":"/api-reference/router/comfy-router/list-the-models-comfy-router-can-run","matchLevel":1},"zh":{"href":"/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run","matchLevel":0},"ja":{"href":"/ja/api-reference/router/comfy-router/list-the-models-comfy-router-can-run","matchLevel":1},"ko":{"href":"/ko/api-reference/router/comfy-router/list-the-models-comfy-router-can-run","matchLevel":1},"es":{"href":"/es/api-reference/router/comfy-router/list-the-models-comfy-router-can-run","matchLevel":1}}},"lastModified":"2026-11-09T12:33:50.274Z","originalFileLocation":"zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run.mdx","locale":"zh"},"headHtml":"\u003cscript data-mint-head>(function(){var d=document.documentElement,p=null;if(!p){try{p=localStorage.getItem(\"isDarkMode\")}catch(e){}if(p==='true')p='dark';else if(p==='false')p='light';if(p!=='light'\u0026\u0026p!=='dark'\u0026\u0026p!=='system')p=\"dark\"}var r=p==='system'?(matchMedia(\"(prefers-color-scheme: dark)\").matches?'dark':'light'):p;d.classList.remove(r==='dark'?'light':'dark');d.classList.add(r);d.setAttribute(\"data-theme-preference\",p)})();\u003c/script>\u003cmeta data-mint-head name=\"description\" content=\"Comfy Router\u0026#x27;s model catalog - one page of the canonical model IDs that POST /v2/models/{provider}/{model} accepts. An SDK calls this on cold start to discover what is runnable, and the model_not_found suggestions come from the same catalog, so an ID listed here that then 404s on invocation would be worse than either failure alone. That agreement is structural rather than a promise: an entry\u0026#x27;s provider and model are the two path segments of the invocation route and reference the same schema components that route\u0026#x27;s path parameters do, and id is those two segments joined by /.\">\u003cmeta data-mint-head name=\"application-name\" content=\"ComfyUI\">\u003cmeta data-mint-head name=\"generator\" content=\"Mintlify\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"192x192\" href=\"/mintlify-assets/_mintlify/favicons/dripart/lS3H5I41bE6LSXEf/_generated/favicon/android-chrome-192x192.png\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"apple-touch-icon\" type=\"image/png\" sizes=\"180x180\" href=\"/mintlify-assets/_mintlify/favicons/dripart/lS3H5I41bE6LSXEf/_generated/favicon/apple-touch-icon.png\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"16x16\" href=\"/mintlify-assets/_mintlify/favicons/dripart/lS3H5I41bE6LSXEf/_generated/favicon/favicon-16x16.png\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"32x32\" href=\"/mintlify-assets/_mintlify/favicons/dripart/lS3H5I41bE6LSXEf/_generated/favicon/favicon-32x32.png\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"shortcut icon\" type=\"image/x-icon\" href=\"/mintlify-assets/_mintlify/favicons/dripart/lS3H5I41bE6LSXEf/_generated/favicon/favicon.ico\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"192x192\" href=\"/mintlify-assets/_mintlify/favicons/dripart/lS3H5I41bE6LSXEf/_generated/favicon-dark/android-chrome-192x192.png\" media=\"(prefers-color-scheme: dark)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"16x16\" href=\"/mintlify-assets/_mintlify/favicons/dripart/lS3H5I41bE6LSXEf/_generated/favicon-dark/favicon-16x16.png\" media=\"(prefers-color-scheme: dark)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"32x32\" href=\"/mintlify-assets/_mintlify/favicons/dripart/lS3H5I41bE6LSXEf/_generated/favicon-dark/favicon-32x32.png\" media=\"(prefers-color-scheme: dark)\">\u003clink data-mint-head rel=\"shortcut icon\" type=\"image/x-icon\" href=\"/mintlify-assets/_mintlify/favicons/dripart/lS3H5I41bE6LSXEf/_generated/favicon-dark/favicon.ico\" media=\"(prefers-color-scheme: dark)\">\u003cmeta data-mint-head property=\"og:title\" content=\"List the models Comfy Router can run. - ComfyUI\">\u003cmeta data-mint-head property=\"og:description\" content=\"Comfy Router\u0026#x27;s model catalog - one page of the canonical model IDs that POST /v2/models/{provider}/{model} accepts. An SDK calls this on cold start to discover what is runnable, and the model_not_found suggestions come from the same catalog, so an ID listed here that then 404s on invocation would be worse than either failure alone. That agreement is structural rather than a promise: an entry\u0026#x27;s provider and model are the two path segments of the invocation route and reference the same schema components that route\u0026#x27;s path parameters do, and id is those two segments joined by /.\">\u003cmeta data-mint-head property=\"og:site_name\" content=\"ComfyUI\">\u003cmeta data-mint-head property=\"og:url\" content=\"https://docs.comfy.org/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run\">\u003cmeta data-mint-head property=\"og:type\" content=\"website\">\u003cmeta data-mint-head property=\"og:image\" content=\"https://dripart.mintlify.site/_mintlify/api/og/v2?division=Comfy+Router\u0026amp;title=List+the+models+Comfy+Router+can+run.\u0026amp;description=Comfy+Router%27s+model+catalog+-+one+page+of+the+canonical+model+IDs+that+%60POST+%2Fv2%2Fmodels%2F%7Bprovider%7D%2F%7Bmodel%7D%60+accepts.+An+SDK+calls+this+on+cold+start+to+discove\u0026amp;theme=b252c39adf72ba65053766a0\u0026amp;signature=3983c304300d9946db3449109d0549a8616ccccdcfb52a3b1fc3c95842a986ed\">\u003cmeta data-mint-head property=\"og:image:width\" content=\"1200\">\u003cmeta data-mint-head property=\"og:image:height\" content=\"630\">\u003cmeta data-mint-head name=\"twitter:card\" content=\"summary_large_image\">\u003cmeta data-mint-head name=\"twitter:title\" content=\"List the models Comfy Router can run. - ComfyUI\">\u003cmeta data-mint-head name=\"twitter:description\" content=\"Comfy Router\u0026#x27;s model catalog - one page of the canonical model IDs that POST /v2/models/{provider}/{model} accepts. An SDK calls this on cold start to discover what is runnable, and the model_not_found suggestions come from the same catalog, so an ID listed here that then 404s on invocation would be worse than either failure alone. That agreement is structural rather than a promise: an entry\u0026#x27;s provider and model are the two path segments of the invocation route and reference the same schema components that route\u0026#x27;s path parameters do, and id is those two segments joined by /.\">\u003cmeta data-mint-head name=\"twitter:image\" content=\"https://dripart.mintlify.site/_mintlify/api/og/v2?division=Comfy+Router\u0026amp;title=List+the+models+Comfy+Router+can+run.\u0026amp;description=Comfy+Router%27s+model+catalog+-+one+page+of+the+canonical+model+IDs+that+%60POST+%2Fv2%2Fmodels%2F%7Bprovider%7D%2F%7Bmodel%7D%60+accepts.+An+SDK+calls+this+on+cold+start+to+discove\u0026amp;theme=b252c39adf72ba65053766a0\u0026amp;signature=3983c304300d9946db3449109d0549a8616ccccdcfb52a3b1fc3c95842a986ed\">\u003cmeta data-mint-head name=\"twitter:image:width\" content=\"1200\">\u003cmeta data-mint-head name=\"twitter:image:height\" content=\"630\">\u003cmeta data-mint-head name=\"msapplication-config\" content=\"/mintlify-assets/_mintlify/favicons/dripart/lS3H5I41bE6LSXEf/_generated/favicon/browserconfig.xml\">\u003cmeta data-mint-head name=\"apple-mobile-web-app-title\" content=\"ComfyUI\">\u003cmeta data-mint-head name=\"msapplication-TileColor\" content=\"#49378B\">\u003cmeta data-mint-head name=\"googlebot\" content=\"max-video-preview:0\">\u003clink data-mint-head rel=\"canonical\" href=\"https://docs.comfy.org/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run\">\u003clink data-mint-head rel=\"alternate\" type=\"application/xml\" href=\"/sitemap.xml\">\u003clink data-mint-head rel=\"alternate\" type=\"text/markdown\" href=\"/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run.md\">\u003clink data-mint-head rel=\"alternate\" hreflang=\"zh\" href=\"https://docs.comfy.org/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run\">\u003clink data-mint-head rel=\"alternate\" hreflang=\"en\" href=\"https://docs.comfy.org/api-reference/router/comfy-router/list-the-models-comfy-router-can-run\">\u003clink data-mint-head rel=\"alternate\" hreflang=\"ja\" href=\"https://docs.comfy.org/ja/api-reference/router/comfy-router/list-the-models-comfy-router-can-run\">\u003clink data-mint-head rel=\"alternate\" hreflang=\"ko\" href=\"https://docs.comfy.org/ko/api-reference/router/comfy-router/list-the-models-comfy-router-can-run\">\u003clink data-mint-head rel=\"alternate\" hreflang=\"es\" href=\"https://docs.comfy.org/es/api-reference/router/comfy-router/list-the-models-comfy-router-can-run\">\u003clink data-mint-head rel=\"alternate\" hreflang=\"x-default\" href=\"https://docs.comfy.org/api-reference/router/comfy-router/list-the-models-comfy-router-can-run\">\u003cscript data-mint-head type=\"application/ld+json\">{\"@context\":\"https://schema.org\",\"@graph\":[{\"@type\":\"Organization\",\"@id\":\"https://docs.comfy.org/#organization\",\"name\":\"ComfyUI\",\"url\":\"https://docs.comfy.org\",\"logo\":{\"@type\":\"ImageObject\",\"url\":\"https://mintcdn.com/dripart/WCFdTagRdhtvk9-o/logo/light.svg?fit=max\u0026auto=format\u0026n=WCFdTagRdhtvk9-o\u0026q=85\u0026s=ec112de87239058f5a931a7e4b594965\"}},{\"@type\":\"WebSite\",\"@id\":\"https://docs.comfy.org#website\",\"name\":\"ComfyUI\",\"url\":\"https://docs.comfy.org\",\"publisher\":{\"@id\":\"https://docs.comfy.org/#organization\"}},{\"@type\":\"WebPage\",\"@id\":\"https://docs.comfy.org/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run#webpage\",\"url\":\"https://docs.comfy.org/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run\",\"name\":\"List the models Comfy Router can run.\",\"description\":\"Comfy Router's model catalog - one page of the canonical model IDs that POST /v2/models/{provider}/{model} accepts. An SDK calls this on cold start to discover what is runnable, and the model_not_found suggestions come from the same catalog, so an ID listed here that then 404s on invocation would be worse than either failure alone. That agreement is structural rather than a promise: an entry's provider and model are the two path segments of the invocation route and reference the same schema components that route's path parameters do, and id is those two segments joined by /.\",\"dateModified\":\"2026-11-09T12:33:50.274Z\",\"isPartOf\":{\"@id\":\"https://docs.comfy.org#website\"},\"breadcrumb\":{\"@id\":\"https://docs.comfy.org/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run#breadcrumb\"}},{\"@type\":\"BreadcrumbList\",\"@id\":\"https://docs.comfy.org/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Comfy Router\",\"item\":\"https://docs.comfy.org/zh/development/comfy-router/quickstart\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"List the models Comfy Router can run.\",\"item\":\"https://docs.comfy.org/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run\"}]},{\"@type\":[\"TechArticle\",\"APIReference\"],\"@id\":\"https://docs.comfy.org/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run#apireference\",\"headline\":\"List the models Comfy Router can run.\",\"name\":\"List the models Comfy Router can run.\",\"description\":\"Comfy Router's model catalog - one page of the canonical model IDs that POST /v2/models/{provider}/{model} accepts. An SDK calls this on cold start to discover what is runnable, and the model_not_found suggestions come from the same catalog, so an ID listed here that then 404s on invocation would be worse than either failure alone. That agreement is structural rather than a promise: an entry's provider and model are the two path segments of the invocation route and reference the same schema components that route's path parameters do, and id is those two segments joined by /.\",\"url\":\"https://docs.comfy.org/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run\",\"mainEntityOfPage\":{\"@id\":\"https://docs.comfy.org/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run#webpage\"},\"dateModified\":\"2026-11-09T12:33:50.274Z\",\"publisher\":{\"@id\":\"https://docs.comfy.org/#organization\"},\"isPartOf\":{\"@id\":\"https://docs.comfy.org#website\"}}]}\u003c/script>\u003cscript data-mint-head type=\"application/ld+json\">{\"@context\":\"https://schema.org\",\"@type\":\"WebSite\",\"name\":\"ComfyUI\",\"creator\":{\"@type\":\"Organization\",\"name\":\"Mintlify\",\"url\":\"https://mintlify.com\"}}\u003c/script>","siteName":"ComfyUI","nav":{"global":null,"languages":[{"language":"en","pages":[{"title":"en","href":"/api-reference/router/comfy-router/list-the-models-comfy-router-can-run"}]},{"language":"zh","footer":{"socials":{"github":"https://github.com/Comfy-Org/ComfyUI","x":"https://x.com/ComfyUI","discord":"https://discord.com/invite/comfyorg","youtube":"https://www.youtube.com/@comfyorg"},"links":[{"header":"资源","items":[{"label":"安装","href":"/zh/installation/system_requirements"},{"label":"教程","href":"/zh/tutorials/basic/text-to-image"},{"label":"开发","href":"/zh/development/overview"}]},{"header":"产品","items":[{"label":"功能","href":"https://www.comfy.org/?utm_source=docs"},{"label":"画廊","href":"https://www.comfy.org/gallery?utm_source=docs"},{"label":"下载","href":"https://www.comfy.org/download?utm_source=docs"}]},{"header":"公司","items":[{"label":"关于","href":"https://www.comfy.org/about?utm_source=docs"},{"label":"招聘","href":"https://www.comfy.org/careers?utm_source=docs"},{"label":"服务条款","href":"https://www.comfy.org/terms-of-service?utm_source=docs"},{"label":"隐私政策","href":"https://www.comfy.org/privacy-policy?utm_source=docs"}]}]},"tabs":[{"tab":"开始使用","pages":[{"title":"开始使用","href":"/zh"}]},{"tab":"内置节点","pages":[{"title":"内置节点","href":"/zh/built-in-nodes/overview"}]},{"tab":"API 开发","pages":[{"group":"使用 Comfy 开发","pages":[{"title":"ComfyUI 开发者指南:SDK、服务器 API 与云端","sidebarTitle":"开发者概览","href":"/zh/development/overview"},{"group":"Comfy Cloud","pages":[{"title":"v1 Cloud API 概述","sidebarTitle":"概述","href":"/zh/development/cloud/overview"},{"title":"Cloud API 参考","icon":"book-open","href":"/zh/development/cloud/api-reference"},{"title":"OpenAPI 规范","icon":"file-lines","openapi":"/openapi-cloud.yaml","href":"/zh/development/cloud/openapi"}]},{"group":"自托管服务器","pages":[{"title":"服务器概述","sidebarTitle":"服务器概览","href":"/zh/development/comfyui-server/comms_overview"},{"title":"ComfyUI Server API 示例:HTTP、WebSocket 与历史记录","sidebarTitle":"API 示例","href":"/zh/development/comfyui-server/api-examples"},{"title":"ComfyUI 服务器路由(HTTP 和 WebSocket API)","sidebarTitle":"HTTP 与 WebSocket 路由","icon":"route","href":"/zh/development/comfyui-server/comms_routes"},{"title":"ComfyUI 服务器消息(WebSocket API)","sidebarTitle":"WebSocket 消息","icon":"comment","href":"/zh/development/comfyui-server/comms_messages"},{"title":"ComfyUI 启动参数:命令行参考","sidebarTitle":"启动参数","icon":"terminal","href":"/zh/development/comfyui-server/startup-flags"},{"title":"自托管选项","sidebarTitle":"部署方式","icon":"server","href":"/zh/development/deploy/self-hosting"}]},{"title":"部署 ComfyUI","sidebarTitle":"概述","icon":"map","href":"/zh/development/deploy/overview"},{"title":"Comfy Cloud 快速开始","sidebarTitle":"快速开始","icon":"cloud","href":"/zh/development/deploy/cloud"},{"title":"Comfy API 和 Router 示例","sidebarTitle":"示例","icon":"code","mode":"wide","href":"/zh/development/examples"}]},{"group":"运行工作流","pages":[{"title":"运行工作流","sidebarTitle":"概述","icon":"play","href":"/zh/development/run-workflows/overview"},{"title":"运行你的第一个工作流","sidebarTitle":"首个工作流","href":"/zh/development/api-development/quickstart"},{"title":"Comfy SDK:ComfyUI 的 Python 与 TypeScript 客户端","sidebarTitle":"SDK","icon":"code","href":"/zh/development/api-development/sdks"},{"title":"自托管 ComfyUI 的 API 代理","sidebarTitle":"API 代理","icon":"plug","href":"/zh/development/comfyui-server/api-proxy"},{"title":"Comfy API:Serverless 部署","icon":"cloud-arrow-up","href":"/zh/development/serverless/overview"},{"title":"获取 API Key","sidebarTitle":"API 密钥","icon":"key","href":"/zh/development/api-development/getting-an-api-key"},{"title":"Comfy API v2 概览","sidebarTitle":"Overview","href":"/zh/api-reference/v2/overview"},{"group":"Comfy API v2 参考文档","pages":[{"group":"assets","pages":[{"title":"Upload an asset (single-call multipart)","deprecated":null,"version":null,"openapi":"/openapi-v2.yaml post /api/v2/assets","href":"/zh/api-reference/v2/assets/upload-an-asset-single-call-multipart"},{"title":"Create or reuse an asset over existing bytes","deprecated":null,"version":null,"openapi":"/openapi-v2.yaml post /api/v2/assets/from-hash","href":"/zh/api-reference/v2/assets/create-or-reuse-an-asset-over-existing-bytes"},{"title":"Existence check for a blob by blake3 hash","deprecated":null,"version":null,"openapi":"/openapi-v2.yaml head /api/v2/assets/by-hash/{hash}","href":"/zh/api-reference/v2/assets/existence-check-for-a-blob-by-blake3-hash"},{"title":"Asset metadata","deprecated":null,"version":null,"openapi":"/openapi-v2.yaml get /api/v2/assets/{id}","href":"/zh/api-reference/v2/assets/asset-metadata"},{"title":"Delete an asset record","deprecated":null,"version":null,"openapi":"/openapi-v2.yaml delete /api/v2/assets/{id}","href":"/zh/api-reference/v2/assets/delete-an-asset-record"},{"title":"Asset bytes","deprecated":null,"version":null,"openapi":"/openapi-v2.yaml get /api/v2/assets/{id}/content","href":"/zh/api-reference/v2/assets/asset-bytes"}]},{"group":"jobs","pages":[{"title":"List jobs, newest first","deprecated":null,"version":null,"openapi":"/openapi-v2.yaml get /api/v2/jobs","href":"/zh/api-reference/v2/jobs/list-jobs-newest-first"},{"title":"Submit a workflow for execution","deprecated":null,"version":null,"openapi":"/openapi-v2.yaml post /api/v2/jobs","href":"/zh/api-reference/v2/jobs/submit-a-workflow-for-execution"},{"title":"Job status (the polling workhorse)","deprecated":null,"version":null,"openapi":"/openapi-v2.yaml get /api/v2/jobs/{id}","href":"/zh/api-reference/v2/jobs/job-status-the-polling-workhorse"},{"title":"The workflow behind a job — authoring version if pinned, executed graph otherwise","deprecated":null,"version":null,"openapi":"/openapi-v2.yaml get /api/v2/jobs/{id}/workflow","href":"/zh/api-reference/v2/jobs/the-workflow-behind-a-job-—-authoring-version-if-pinned-executed-graph-otherwise"},{"title":"What the run printed","deprecated":null,"version":null,"openapi":"/openapi-v2.yaml get /api/v2/jobs/{id}/logs","href":"/zh/api-reference/v2/jobs/what-the-run-printed"},{"title":"Live event stream (SSE)","deprecated":null,"version":null,"openapi":"/openapi-v2.yaml get /api/v2/jobs/{id}/events","href":"/zh/api-reference/v2/jobs/live-event-stream-sse"},{"title":"Request cancellation","deprecated":null,"version":null,"openapi":"/openapi-v2.yaml post /api/v2/jobs/{id}/cancel","href":"/zh/api-reference/v2/jobs/request-cancellation"}]}]},{"title":"设计说明","sidebarTitle":"设计说明","icon":"compass","href":"/zh/development/api-development/sdks-design"},{"title":"工作流 API 格式","icon":"file-code","href":"/zh/development/api-development/workflow-api-format"},{"title":"工作流元数据:ComfyUI 输出文件中嵌入的工作流","sidebarTitle":"工作流元数据","icon":"file-code","href":"/zh/development/api-development/workflow-metadata"},{"title":"ComfyUI 账户 API 密钥集成","sidebarTitle":"API 密钥集成","icon":"puzzle-piece","href":"/zh/development/comfyui-server/api-key-integration"},{"group":"Comfy API Reference","pages":[{"title":"Comfy API 版本:v2 与已弃用的 v1 Cloud API","sidebarTitle":"概述","href":"/zh/development/comfy-api/overview"}]}]},{"group":"Comfy Router","pages":[{"title":"Comfy Router 快速入门","sidebarTitle":"快速入门","href":"/zh/development/comfy-router/quickstart"},{"title":"Comfy Router API 概览","sidebarTitle":"概览","href":"/zh/development/comfy-router/api"},{"title":"发现 Comfy Router 模型","sidebarTitle":"发现模型","href":"/zh/development/comfy-router/discover-models"},{"title":"Comfy Router 模型 schema 与结果","sidebarTitle":"Schema 与结果","href":"/zh/development/comfy-router/schemas"},{"title":"Comfy Router 请求排队","sidebarTitle":"排队的请求","href":"/zh/development/comfy-router/queue"},{"title":"Comfy Router 错误与重试","sidebarTitle":"错误与重试","href":"/zh/development/comfy-router/errors"},{"title":"Comfy Router 计费","sidebarTitle":"计费","href":"/zh/development/comfy-router/billing"},{"title":"将 Comfy Router 与 LiteLLM 配合使用","sidebarTitle":"LiteLLM","href":"/zh/development/comfy-router/litellm"},{"title":"Comfy Router 实时语音会话","sidebarTitle":"实时语音","href":"/zh/development/comfy-router/realtime"},{"title":"Comfy Router 请求头","sidebarTitle":"请求头","href":"/zh/development/comfy-router/headers"},{"title":"Comfy Router API 参考","sidebarTitle":"API 参考","href":"/zh/development/comfy-router/reference"},{"title":"Comfy Router 的能力与限制","sidebarTitle":"能力与限制","description":"选择 Comfy Router 还是合作伙伴代理,为长时间运行的调用做好规划,并了解恢复、速率限制与资产存储。","href":"/zh/development/comfy-router/limitations"},{"group":"Comfy Router API 参考文档","pages":[{"group":"Comfy Router","pages":[{"title":"List the models Comfy Router can run.","deprecated":null,"version":null,"openapi":"/router-openapi.yaml get /v2/models","href":"/zh/api-reference/router/comfy-router/list-the-models-comfy-router-can-run"},{"title":"Read one partner model's catalog entry by canonical model ID.","description":"Per-model detail for a single Comfy Router model, so a caller can check one model without walking the whole paginated catalog. The SDKs use it to look a model up immediately before invoking it.","deprecated":null,"version":null,"openapi":"/router-openapi.yaml get /v2/models/{provider}/{model}","href":"/zh/api-reference/router/comfy-router/read-one-partner-models-catalog-entry-by-canonical-model-id"},{"title":"Run a partner model synchronously by canonical model ID.","deprecated":null,"version":null,"openapi":"/router-openapi.yaml post /v2/models/{provider}/{model}","href":"/zh/api-reference/router/comfy-router/run-a-partner-model-synchronously-by-canonical-model-id"},{"title":"Price a model call before running it, without running it.","deprecated":null,"version":null,"openapi":"/router-openapi.yaml post /v2/models/{provider}/{model}/estimate","href":"/zh/api-reference/router/comfy-router/price-a-model-call-before-running-it-without-running-it"},{"title":"Read one partner model's input and output schemas as an OpenAPI document.","deprecated":null,"version":null,"openapi":"/router-openapi.yaml get /v2/models/{provider}/{model}/openapi.json","href":"/zh/api-reference/router/comfy-router/read-one-partner-models-input-and-output-schemas-as-an-openapi-document"},{"title":"Submit a partner model run to the queue and return immediately.","deprecated":null,"version":null,"openapi":"/router-openapi.yaml post /v2/models/{provider}/{model}/requests","href":"/zh/api-reference/router/comfy-router/submit-a-partner-model-run-to-the-queue-and-return-immediately"},{"title":"Collect the result of one submitted request.","deprecated":null,"version":null,"openapi":"/router-openapi.yaml get /v2/models/{provider}/{model}/requests/{request_id}","href":"/zh/api-reference/router/comfy-router/collect-the-result-of-one-submitted-request"},{"title":"Ask for one submitted request to be cancelled.","deprecated":null,"version":null,"openapi":"/router-openapi.yaml put /v2/models/{provider}/{model}/requests/{request_id}/cancel","href":"/zh/api-reference/router/comfy-router/ask-for-one-submitted-request-to-be-cancelled"},{"title":"Read the queue state of one submitted request.","deprecated":null,"version":null,"openapi":"/router-openapi.yaml get /v2/models/{provider}/{model}/requests/{request_id}/status","href":"/zh/api-reference/router/comfy-router/read-the-queue-state-of-one-submitted-request"}]}]},{"group":"Models","pages":[{"title":"Comfy Router 模型","sidebarTitle":"所有模型","href":"/zh/development/comfy-router/models"},{"title":"Comfy Router 定价","sidebarTitle":"定价","mode":"wide","href":"/zh/development/comfy-router/pricing"},{"title":"Comfy Router 服务提供商","sidebarTitle":"服务提供商","href":"/zh/development/comfy-router/providers"},{"group":"Anthropic","pages":[{"title":"将 Claude Fable 5.1 与 Comfy Router 配合使用","sidebarTitle":"Claude Fable 5.1","href":"/zh/development/comfy-router/models/anthropic/claude-fable-5-1/code"},{"title":"将 Claude Fable 5 与 Comfy Router 配合使用","sidebarTitle":"Claude Fable 5","href":"/zh/development/comfy-router/models/anthropic/claude-fable-5/code"},{"title":"将 Claude Haiku 4.5 20251001 与 Comfy Router 配合使用","sidebarTitle":"Claude Haiku 4.5 20251001","href":"/zh/development/comfy-router/models/anthropic/claude-haiku-4-5-20251001/code"},{"title":"将 Claude Opus 4.6 与 Comfy Router 配合使用","sidebarTitle":"Claude Opus 4.6","href":"/zh/development/comfy-router/models/anthropic/claude-opus-4-6/code"},{"title":"将 Claude Opus 4.7 与 Comfy Router 配合使用","sidebarTitle":"Claude Opus 4.7","href":"/zh/development/comfy-router/models/anthropic/claude-opus-4-7/code"},{"title":"配合 Comfy Router 使用 Claude Opus 4.8","sidebarTitle":"Claude Opus 4.8","href":"/zh/development/comfy-router/models/anthropic/claude-opus-4-8/code"},{"title":"使用 Claude Opus 5.5 搭配 Comfy Router","sidebarTitle":"Claude Opus 5.5","href":"/zh/development/comfy-router/models/anthropic/claude-opus-5-5/code"},{"title":"将 Claude Opus 5 与 Comfy Router 配合使用","sidebarTitle":"Claude Opus 5","href":"/zh/development/comfy-router/models/anthropic/claude-opus-5/code"},{"title":"将 Claude Sonnet 4.5 20250929 与 Comfy Router 配合使用","sidebarTitle":"Claude Sonnet 4.5 20250929","href":"/zh/development/comfy-router/models/anthropic/claude-sonnet-4-5-20250929/code"},{"title":"使用 Comfy Router 调用 Claude Sonnet 4.6","sidebarTitle":"Claude Sonnet 4.6","href":"/zh/development/comfy-router/models/anthropic/claude-sonnet-4-6/code"},{"title":"结合 Comfy Router 使用 Claude Sonnet 5.5","sidebarTitle":"Claude Sonnet 5.5","href":"/zh/development/comfy-router/models/anthropic/claude-sonnet-5-5/code"},{"title":"将 Claude Sonnet 5 与 Comfy Router 配合使用","sidebarTitle":"Claude Sonnet 5","href":"/zh/development/comfy-router/models/anthropic/claude-sonnet-5/code"}]},{"group":"Beeble","pages":[{"title":"将 SwitchX 与 Comfy Router 配合使用","sidebarTitle":"SwitchX","href":"/zh/development/comfy-router/models/beeble/switchx/code"}]},{"group":"Black Forest Labs","pages":[{"title":"搭配 Comfy Router 使用 Erase V1","sidebarTitle":"Erase V1","href":"/zh/development/comfy-router/models/black-forest-labs/erase-v1/code"},{"title":"使用 Flux 1.1 Pro Ultra Image 与 Comfy Router","sidebarTitle":"Flux 1.1 Pro Ultra Image","href":"/zh/development/comfy-router/models/black-forest-labs/flux-1-1-pro-ultra-image/code"},{"title":"通过 Comfy Router 使用 FLUX.1 Kontext","sidebarTitle":"FLUX.1 Kontext","href":"/zh/development/comfy-router/models/black-forest-labs/flux-1-kontext/code"},{"title":"使用 Comfy Router 调用 FLUX 2 Max","sidebarTitle":"FLUX 2 Max","href":"/zh/development/comfy-router/models/black-forest-labs/flux-2-max/code"},{"title":"将 FLUX 2 Pro 与 Comfy Router 配合使用","sidebarTitle":"FLUX 2 Pro","href":"/zh/development/comfy-router/models/black-forest-labs/flux-2-pro/code"},{"title":"配合 Comfy Router 使用 FLUX 3 Image","sidebarTitle":"FLUX 3 Image","href":"/zh/development/comfy-router/models/black-forest-labs/flux-3-image/code"},{"title":"使用 Comfy Router 调用 FLUX 3 Video","sidebarTitle":"FLUX 3 Video","href":"/zh/development/comfy-router/models/black-forest-labs/flux-3-video/code"},{"title":"使用 Comfy Router 调用 FLUX Pro 1.0 Expand","sidebarTitle":"FLUX Pro 1.0 Expand","href":"/zh/development/comfy-router/models/black-forest-labs/flux-pro-1-0-expand/code"},{"title":"使用 Comfy Router 调用 FLUX Pro 1.0 Fill","sidebarTitle":"FLUX Pro 1.0 Fill","href":"/zh/development/comfy-router/models/black-forest-labs/flux-pro-1-0-fill/code"},{"title":"通过 Comfy Router 使用 FLUX Video Upscale","sidebarTitle":"FLUX Video Upscale","href":"/zh/development/comfy-router/models/black-forest-labs/flux-video-upscale/code"},{"title":"将 Video Edit V1 与 Comfy Router 配合使用","sidebarTitle":"Video Edit V1","href":"/zh/development/comfy-router/models/black-forest-labs/video-edit-v1/code"},{"title":"将 VTO V1 与 Comfy Router 搭配使用","sidebarTitle":"VTO V1","href":"/zh/development/comfy-router/models/black-forest-labs/vto-v1/code"}]},{"group":"Bria","pages":[{"title":"将 Fibo 与 Comfy Router 配合使用","sidebarTitle":"Fibo","href":"/zh/development/comfy-router/models/bria/fibo/code"},{"title":"将 Image Edit Add Object By Text 与 Comfy Router 配合使用","sidebarTitle":"Image Edit Add Object By Text","href":"/zh/development/comfy-router/models/bria/image-edit-add-object-by-text/code"},{"title":"结合 Comfy Router 使用 Image Edit Erase By Text","sidebarTitle":"Image Edit Erase By Text","href":"/zh/development/comfy-router/models/bria/image-edit-erase-by-text/code"},{"title":"配合 Comfy Router 使用 Image Edit Erase Foreground","sidebarTitle":"Image Edit Erase Foreground","href":"/zh/development/comfy-router/models/bria/image-edit-erase-foreground/code"},{"title":"在 Comfy Router 中使用 Image Edit Erase","sidebarTitle":"Image Edit Erase","href":"/zh/development/comfy-router/models/bria/image-edit-erase/code"},{"title":"通过 Comfy Router 使用 Image Edit Expand","sidebarTitle":"Image Edit Expand","href":"/zh/development/comfy-router/models/bria/image-edit-expand/code"},{"title":"将 Image Edit Gen Fill 与 Comfy Router 配合使用","sidebarTitle":"Image Edit Gen Fill","href":"/zh/development/comfy-router/models/bria/image-edit-gen-fill/code"},{"title":"搭配 Comfy Router 使用 Image Edit Increase Resolution","sidebarTitle":"Image Edit Increase Resolution","href":"/zh/development/comfy-router/models/bria/image-edit-increase-resolution/code"},{"title":"使用 Comfy Router 调用 Image Edit Product Holding","sidebarTitle":"Image Edit Product Holding","href":"/zh/development/comfy-router/models/bria/image-edit-product-holding/code"},{"title":"结合 Comfy Router 使用 Image Edit Product Virtual Tryon","sidebarTitle":"Image Edit Product Virtual Tryon","href":"/zh/development/comfy-router/models/bria/image-edit-product-virtual-tryon/code"},{"title":"将 Image Edit Relight 与 Comfy Router 配合使用","sidebarTitle":"Image Edit Relight","href":"/zh/development/comfy-router/models/bria/image-edit-relight/code"},{"title":"使用 Comfy Router 调用 Image Edit Remove Background Guided","sidebarTitle":"Image Edit Remove Background Guided","href":"/zh/development/comfy-router/models/bria/image-edit-remove-background-guided/code"},{"title":"通过 Comfy Router 使用 Image Edit Remove Background","sidebarTitle":"Image Edit Remove Background","href":"/zh/development/comfy-router/models/bria/image-edit-remove-background/code"},{"title":"使用 Comfy Router 调用 Image Edit Replace Background","sidebarTitle":"Image Edit Replace Background","href":"/zh/development/comfy-router/models/bria/image-edit-replace-background/code"},{"title":"配合 Comfy Router 使用 Image Edit Replace Object By Text","sidebarTitle":"Image Edit Replace Object By Text","href":"/zh/development/comfy-router/models/bria/image-edit-replace-object-by-text/code"},{"title":"通过 Comfy Router 使用 Image Edit Reseason","sidebarTitle":"Image Edit Reseason","href":"/zh/development/comfy-router/models/bria/image-edit-reseason/code"},{"title":"在 Comfy Router 中使用 Image Edit Restore","sidebarTitle":"Image Edit Restore","href":"/zh/development/comfy-router/models/bria/image-edit-restore/code"},{"title":"配合 Comfy Router 使用图像","sidebarTitle":"图像","href":"/zh/development/comfy-router/models/bria/image/code"},{"title":"将 Video Edit Erase 与 Comfy Router 搭配使用","sidebarTitle":"Video Edit Erase","href":"/zh/development/comfy-router/models/bria/video-edit-erase/code"},{"title":"配合 Comfy Router 使用 Video Edit Green Screen","sidebarTitle":"Video Edit Green Screen","href":"/zh/development/comfy-router/models/bria/video-edit-green-screen/code"},{"title":"使用 Comfy Router 进行视频编辑移除背景","sidebarTitle":"Video Edit Remove Background","href":"/zh/development/comfy-router/models/bria/video-edit-remove-background/code"},{"title":"使用 Comfy Router 进行视频编辑替换背景","sidebarTitle":"Video Edit Replace Background","href":"/zh/development/comfy-router/models/bria/video-edit-replace-background/code"}]},{"group":"BytePlus","pages":[{"title":"使用 Dreamina Seedance 2.0 260128 与 Comfy Router","sidebarTitle":"Dreamina Seedance 2.0 260128","href":"/zh/development/comfy-router/models/byteplus/dreamina-seedance-2-0-260128/code"},{"title":"使用 Comfy Router 调用 Dreamina Seedance 2.0 Fast 260128","sidebarTitle":"Dreamina Seedance 2.0 Fast 260128","href":"/zh/development/comfy-router/models/byteplus/dreamina-seedance-2-0-fast-260128/code"},{"title":"通过 Comfy Router 使用 Dreamina Seedance 2.0 Mini","sidebarTitle":"Dreamina Seedance 2.0 Mini","href":"/zh/development/comfy-router/models/byteplus/dreamina-seedance-2-0-mini/code"},{"title":"使用 Dreamina Seedance 2.5 260628 与 Comfy Router","sidebarTitle":"Dreamina Seedance 2.5 260628","href":"/zh/development/comfy-router/models/byteplus/dreamina-seedance-2-5-260628/code"},{"title":"将 Seed 2.0 Lite 260228 与 Comfy Router 配合使用","sidebarTitle":"Seed 2.0 Lite 260228","href":"/zh/development/comfy-router/models/byteplus/seed-2-0-lite-260228/code"},{"title":"将 Seed 2.0 Mini 260215 与 Comfy Router 配合使用","sidebarTitle":"Seed 2.0 Mini 260215","href":"/zh/development/comfy-router/models/byteplus/seed-2-0-mini-260215/code"},{"title":"将 Seed 2.0 Pro 260328 与 Comfy Router 搭配使用","sidebarTitle":"Seed 2.0 Pro 260328","href":"/zh/development/comfy-router/models/byteplus/seed-2-0-pro-260328/code"},{"title":"将 Seed Audio 1.0 Multilingual 与 Comfy Router 配合使用","sidebarTitle":"Seed Audio 1.0 Multilingual","href":"/zh/development/comfy-router/models/byteplus/seed-audio-1-0-multilingual/code"},{"title":"结合 Comfy Router 使用 Seed Audio 1.0","sidebarTitle":"Seed Audio 1.0","href":"/zh/development/comfy-router/models/byteplus/seed-audio-1-0/code"},{"title":"将 Seedance 1.0 Pro 250528 与 Comfy Router 配合使用","sidebarTitle":"Seedance 1.0 Pro 250528","href":"/zh/development/comfy-router/models/byteplus/seedance-1-0-pro-250528/code"},{"title":"使用 Comfy Router 调用 Seedance 1.0 Pro Fast 251015","sidebarTitle":"Seedance 1.0 Pro Fast 251015","href":"/zh/development/comfy-router/models/byteplus/seedance-1-0-pro-fast-251015/code"},{"title":"通过 Comfy Router 使用 Seedance 1.5 Pro 251215","sidebarTitle":"Seedance 1.5 Pro 251215","href":"/zh/development/comfy-router/models/byteplus/seedance-1-5-pro-251215/code"},{"title":"将 Seedream 4.0 250828 与 Comfy Router 配合使用","sidebarTitle":"Seedream 4.0 250828","href":"/zh/development/comfy-router/models/byteplus/seedream-4-0-250828/code"},{"title":"将 Seedream 4.5 251128 与 Comfy Router 配合使用","sidebarTitle":"Seedream 4.5 251128","href":"/zh/development/comfy-router/models/byteplus/seedream-4-5-251128/code"},{"title":"结合 Comfy Router 使用 Seedream 5.0 260128","sidebarTitle":"Seedream 5.0 260128","href":"/zh/development/comfy-router/models/byteplus/seedream-5-0-260128/code"},{"title":"将 Seedream 5.0 Flash 260915 与 Comfy Router 配合使用","sidebarTitle":"Seedream 5.0 Flash 260915","href":"/zh/development/comfy-router/models/byteplus/seedream-5-0-flash-260915/code"},{"title":"使用 Comfy Router 调用 Seedream 5.0 Pro 260628","sidebarTitle":"Seedream 5.0 Pro 260628","href":"/zh/development/comfy-router/models/byteplus/seedream-5-0-pro-260628/code"}]},{"group":"Elevenlabs","pages":[{"title":"使用 Comfy Router 调用 Eleven Multilingual V2","sidebarTitle":"Eleven Multilingual V2","href":"/zh/development/comfy-router/models/elevenlabs/eleven-multilingual-v2/code"},{"title":"将 Eleven Sfx V2 与 Comfy Router 配合使用","sidebarTitle":"Eleven Sfx V2","href":"/zh/development/comfy-router/models/elevenlabs/eleven-sfx-v2/code"},{"title":"将 Eleven V3 与 Comfy Router 配合使用","sidebarTitle":"Eleven V3","href":"/zh/development/comfy-router/models/elevenlabs/eleven-v3/code"},{"title":"将 Eleven V4 Turbo 与 Comfy Router 搭配使用","sidebarTitle":"Eleven V4 Turbo","href":"/zh/development/comfy-router/models/elevenlabs/eleven-v4-turbo/code"},{"title":"通过 Comfy Router 使用 Eleven V4","sidebarTitle":"Eleven V4","href":"/zh/development/comfy-router/models/elevenlabs/eleven-v4/code"}]},{"group":"fal","pages":[{"title":"使用 Comfy Router 调用 Crystal Upscaler","sidebarTitle":"Crystal Upscaler","href":"/zh/development/comfy-router/models/fal/crystal-upscaler/code"},{"title":"将 Crystal Video Upscaler 与 Comfy Router 配合使用","sidebarTitle":"Crystal Video Upscaler","href":"/zh/development/comfy-router/models/fal/crystal-video-upscaler/code"},{"title":"通过 Comfy Router 使用 FLUX 3 Max","sidebarTitle":"FLUX 3 Max","href":"/zh/development/comfy-router/models/fal/flux-3-max/code"},{"title":"将 H3 Max Turbo 与 Comfy Router 搭配使用","sidebarTitle":"H3 Max Turbo","href":"/zh/development/comfy-router/models/fal/h3-max-turbo/code"},{"title":"在 Comfy Router 中使用 H3 Max","sidebarTitle":"H3 Max","href":"/zh/development/comfy-router/models/fal/h3-max/code"},{"title":"将 Patina 与 Comfy Router 配合使用","sidebarTitle":"Patina","href":"/zh/development/comfy-router/models/fal/patina/code"}]},{"group":"Freepik","pages":[{"title":"将 AI Image Upscaler Precision V2 与 Comfy Router 配合使用","sidebarTitle":"AI Image Upscaler Precision V2","href":"/zh/development/comfy-router/models/freepik/ai-image-upscaler-precision-v2/code"},{"title":"将 AI Skin Enhancer Creative 与 Comfy Router 配合使用","sidebarTitle":"AI Skin Enhancer Creative","href":"/zh/development/comfy-router/models/freepik/ai-skin-enhancer-creative/code"},{"title":"将 AI Skin Enhancer Faithful 与 Comfy Router 配合使用","sidebarTitle":"AI Skin Enhancer Faithful","href":"/zh/development/comfy-router/models/freepik/ai-skin-enhancer-faithful/code"},{"title":"搭配 Comfy Router 使用 AI Skin Enhancer Flexible","sidebarTitle":"AI Skin Enhancer Flexible","href":"/zh/development/comfy-router/models/freepik/ai-skin-enhancer-flexible/code"}]},{"group":"Gemini Interactions","pages":[{"title":"搭配 Comfy Router 使用 Gemini Omni 1.1 Flash","sidebarTitle":"Gemini Omni 1.1 Flash","href":"/zh/development/comfy-router/models/gemini-interactions/gemini-omni-1-1-flash/code"},{"title":"使用 Gemini Omni 1.2 Flash 配合 Comfy Router","sidebarTitle":"Gemini Omni 1.2 Flash","href":"/zh/development/comfy-router/models/gemini-interactions/gemini-omni-1-2-flash/code"},{"title":"使用 Comfy Router 调用 Gemini Omni Flash Preview","sidebarTitle":"Gemini Omni Flash Preview","href":"/zh/development/comfy-router/models/gemini-interactions/gemini-omni-flash-preview/code"}]},{"group":"Google","pages":[{"title":"使用 Gemini 2.5 Flash Image 与 Comfy Router","sidebarTitle":"Gemini 2.5 Flash Image","href":"/zh/development/comfy-router/models/google/gemini-2-5-flash-image/code"},{"title":"使用 Gemini 3.1 Flash Lite 与 Comfy Router","sidebarTitle":"Gemini 3.1 Flash Lite","href":"/zh/development/comfy-router/models/google/gemini-3-1-flash-lite/code"},{"title":"使用 Gemini 3.7 Flash 搭配 Comfy Router","sidebarTitle":"Gemini 3.7 Flash","href":"/zh/development/comfy-router/models/google/gemini-3-7-flash/code"},{"title":"搭配 Comfy Router 使用 Gemini 3.8 Flash","sidebarTitle":"Gemini 3.8 Flash","href":"/zh/development/comfy-router/models/google/gemini-3-8-flash/code"},{"title":"将 Gemini Nano Banana 2.1 与 Comfy Router 配合使用","sidebarTitle":"Gemini Nano Banana 2.1","href":"/zh/development/comfy-router/models/google/gemini-nano-banana-2-1/code"},{"title":"通过 Comfy Router 使用 Google Gemini","sidebarTitle":"Google Gemini","href":"/zh/development/comfy-router/models/google/gemini/code"},{"title":"将 Nano Banana 2 Lite 与 Comfy Router 配合使用","sidebarTitle":"Nano Banana 2 Lite","href":"/zh/development/comfy-router/models/google/nano-banana-2-lite/code"},{"title":"将 Nano Banana 2 与 Comfy Router 配合使用","sidebarTitle":"Nano Banana 2","href":"/zh/development/comfy-router/models/google/nano-banana-2/code"},{"title":"搭配 Comfy Router 使用 Nano Banana Pro","sidebarTitle":"Nano Banana Pro","href":"/zh/development/comfy-router/models/google/nano-banana-pro/code"}]},{"group":"HeyGen","pages":[{"title":"将 Starfish 与 Comfy Router 配合使用","sidebarTitle":"Starfish","href":"/zh/development/comfy-router/models/heygen/starfish/code"}]},{"group":"Higgsfield","pages":[{"title":"Higgsfield Genjutsu Motion Transfer V1.0:通过 Comfy Router 调用","sidebarTitle":"Higgsfield Genjutsu Motion Transfer V1.0","href":"/zh/development/comfy-router/models/higgsfield/higgsfield-genjutsu-motion-transfer-v1-0/code"},{"title":"将 Higgsfield Kling 3.4k 与 Comfy Router 搭配使用","sidebarTitle":"Higgsfield Kling 3.4k","href":"/zh/development/comfy-router/models/higgsfield/higgsfield-kling-3-4k/code"},{"title":"将 Higgsfield Kling 3 Pro 与 Comfy Router 配合使用","sidebarTitle":"Higgsfield Kling 3 Pro","href":"/zh/development/comfy-router/models/higgsfield/higgsfield-kling-3-pro/code"}]},{"group":"Ideogram","pages":[{"title":"将 Ideogram 4.5 与 Comfy Router 配合使用","sidebarTitle":"Ideogram 4.5","href":"/zh/development/comfy-router/models/ideogram/ideogram-4-5/code"},{"title":"将 Ideogram V3 与 Comfy Router 配合使用","sidebarTitle":"Ideogram V3","href":"/zh/development/comfy-router/models/ideogram/ideogram-v3/code"},{"title":"结合 Comfy Router 使用 Ideogram 4.0","sidebarTitle":"Ideogram 4.0","href":"/zh/development/comfy-router/models/ideogram/ideogram-v4/code"},{"title":"配合 Comfy Router 使用 P Image Ideogram","sidebarTitle":"P Image Ideogram","href":"/zh/development/comfy-router/models/ideogram/p-image-ideogram/code"}]},{"group":"Kling","pages":[{"title":"通过 Comfy Router 使用 Kling 3.0 Turbo","sidebarTitle":"Kling 3.0 Turbo","href":"/zh/development/comfy-router/models/kling/kling-3-0-turbo/code"},{"title":"通过 Comfy Router 使用 Kling 4.0","sidebarTitle":"Kling 4.0","href":"/zh/development/comfy-router/models/kling/kling-4-0/code"},{"title":"使用 Comfy Router 调用 Kling Image O1","sidebarTitle":"Kling Image O1","href":"/zh/development/comfy-router/models/kling/kling-image-o1/code"},{"title":"将 Kling V2.5 Turbo 与 Comfy Router 搭配使用","sidebarTitle":"Kling V2.5 Turbo","href":"/zh/development/comfy-router/models/kling/kling-v2-5-turbo/code"},{"title":"通过 Comfy Router 使用 Kling V2.6","sidebarTitle":"Kling V2.6","href":"/zh/development/comfy-router/models/kling/kling-v2-6/code"},{"title":"通过 Comfy Router 使用 Kling V3 Omni","sidebarTitle":"Kling V3 Omni","href":"/zh/development/comfy-router/models/kling/kling-v3-omni/code"},{"title":"通过 Comfy Router 使用 Kling V3","sidebarTitle":"Kling V3","href":"/zh/development/comfy-router/models/kling/kling-v3/code"},{"title":"将 Kling Video O1 与 Comfy Router 配合使用","sidebarTitle":"Kling Video O1","href":"/zh/development/comfy-router/models/kling/kling-video-o1/code"},{"title":"配合 Comfy Router 使用 Videos Avatar Image 2video","sidebarTitle":"Videos Avatar Image 2video","href":"/zh/development/comfy-router/models/kling/videos-avatar-image2video/code"},{"title":"通过 Comfy Router 使用视频唇形同步","sidebarTitle":"视频唇形同步","href":"/zh/development/comfy-router/models/kling/videos-lip-sync/code"}]},{"group":"Krea","pages":[{"title":"使用 Comfy Router 调用 Krea 2 Large","sidebarTitle":"Krea 2 Large","href":"/zh/development/comfy-router/models/krea/krea-2-large/code"},{"title":"将 Krea 2 Medium Turbo 与 Comfy Router 搭配使用","sidebarTitle":"Krea 2 Medium Turbo","href":"/zh/development/comfy-router/models/krea/krea-2-medium-turbo/code"},{"title":"配合 Comfy Router 使用 Krea 2 Medium","sidebarTitle":"Krea 2 Medium","href":"/zh/development/comfy-router/models/krea/krea-2-medium/code"},{"title":"在 Comfy Router 中使用 Krea 2","sidebarTitle":"Krea 2","href":"/zh/development/comfy-router/models/krea/krea-2/code"},{"title":"通过 Comfy Router 使用 Krea 3","sidebarTitle":"Krea 3","href":"/zh/development/comfy-router/models/krea/krea-3/code"}]},{"group":"LTX","pages":[{"title":"通过 Comfy Router 使用 LTX 2.5 Fast","sidebarTitle":"LTX 2.5 Fast","href":"/zh/development/comfy-router/models/ltx/ltx-2-5-fast/code"},{"title":"使用 Comfy Router 调用 LTX 2.5 Pro","sidebarTitle":"LTX 2.5 Pro","href":"/zh/development/comfy-router/models/ltx/ltx-2-5-pro/code"}]},{"group":"Luma","pages":[{"title":"将 Photon 1 与 Comfy Router 配合使用","sidebarTitle":"Photon 1","href":"/zh/development/comfy-router/models/luma/photon-1/code"},{"title":"将 Photon Flash 1 与 Comfy Router 配合使用","sidebarTitle":"Photon Flash 1","href":"/zh/development/comfy-router/models/luma/photon-flash-1/code"}]},{"group":"Luma 2","pages":[{"title":"搭配 Comfy Router 使用 Ray 3.2","sidebarTitle":"Ray 3.2","href":"/zh/development/comfy-router/models/luma_2/ray-3-2/code"},{"title":"搭配 Comfy Router 使用 Uni 1 Max","sidebarTitle":"Uni 1 Max","href":"/zh/development/comfy-router/models/luma_2/uni-1-max/code"},{"title":"将 Uni 1 与 Comfy Router 配合使用","sidebarTitle":"Uni 1","href":"/zh/development/comfy-router/models/luma_2/uni-1/code"}]},{"group":"Meshy","pages":[{"title":"将 Animations 与 Comfy Router 配合使用","sidebarTitle":"Animations","href":"/zh/development/comfy-router/models/meshy/animations/code"},{"title":"搭配 Comfy Router 使用 Meshy 6","sidebarTitle":"Meshy 6","href":"/zh/development/comfy-router/models/meshy/meshy-6/code"},{"title":"将 Meshy 7.1 与 Comfy Router 配合使用","sidebarTitle":"Meshy 7.1","href":"/zh/development/comfy-router/models/meshy/meshy-7-1/code"},{"title":"将 Meshy 7 与 Comfy Router 配合使用","sidebarTitle":"Meshy 7","href":"/zh/development/comfy-router/models/meshy/meshy-7/code"},{"title":"将 Remesh 与 Comfy Router 配合使用","sidebarTitle":"Remesh","href":"/zh/development/comfy-router/models/meshy/remesh/code"},{"title":"通过 Comfy Router 使用 Rigging","sidebarTitle":"Rigging","href":"/zh/development/comfy-router/models/meshy/rigging/code"}]},{"group":"MiniMax","pages":[{"title":"配合 Comfy Router 使用 MiniMax H3","sidebarTitle":"MiniMax H3","href":"/zh/development/comfy-router/models/minimax/minimax-h3/code"}]},{"group":"Moonvalley","pages":[{"title":"使用 Comfy Router 进行图生视频","sidebarTitle":"图生视频","href":"/zh/development/comfy-router/models/moonvalley/image-to-video/code"},{"title":"使用 Comfy Router 进行文生视频","sidebarTitle":"文生视频","href":"/zh/development/comfy-router/models/moonvalley/text-to-video/code"},{"title":"通过 Comfy Router 使用视频转视频","sidebarTitle":"视频转视频","href":"/zh/development/comfy-router/models/moonvalley/video-to-video/code"}]},{"group":"OpenAI","pages":[{"title":"将 GPT 4.1 Mini 与 Comfy Router 配合使用","sidebarTitle":"GPT 4.1 Mini","href":"/zh/development/comfy-router/models/openai/gpt-4-1-mini/code"},{"title":"将 GPT 4.1 与 Comfy Router 配合使用","sidebarTitle":"GPT 4.1","href":"/zh/development/comfy-router/models/openai/gpt-4-1/code"},{"title":"将 GPT 4o 与 Comfy Router 配合使用","sidebarTitle":"GPT 4o","href":"/zh/development/comfy-router/models/openai/gpt-4o/code"},{"title":"使用 GPT 5.5 Pro 搭配 Comfy Router","sidebarTitle":"GPT 5.5 Pro","href":"/zh/development/comfy-router/models/openai/gpt-5-5-pro/code"},{"title":"将 GPT 5.5 与 Comfy Router 配合使用","sidebarTitle":"GPT 5.5","href":"/zh/development/comfy-router/models/openai/gpt-5-5/code"},{"title":"将 GPT 5.6 Luna 与 Comfy Router 配合使用","sidebarTitle":"GPT 5.6 Luna","href":"/zh/development/comfy-router/models/openai/gpt-5-6-luna/code"},{"title":"通过 Comfy Router 使用 GPT 5.6 Sol","sidebarTitle":"GPT 5.6 Sol","href":"/zh/development/comfy-router/models/openai/gpt-5-6-sol/code"},{"title":"将 GPT 5.6 Terra 与 Comfy Router 配合使用","sidebarTitle":"GPT 5.6 Terra","href":"/zh/development/comfy-router/models/openai/gpt-5-6-terra/code"},{"title":"在 Comfy Router 中使用 GPT 5 Mini","sidebarTitle":"GPT 5 Mini","href":"/zh/development/comfy-router/models/openai/gpt-5-mini/code"},{"title":"将 GPT 5 Nano 与 Comfy Router 结合使用","sidebarTitle":"GPT 5 Nano","href":"/zh/development/comfy-router/models/openai/gpt-5-nano/code"},{"title":"将 GPT 5 与 Comfy Router 搭配使用","sidebarTitle":"GPT 5","href":"/zh/development/comfy-router/models/openai/gpt-5/code"},{"title":"使用 GPT 6.1 Sol 与 Comfy Router","sidebarTitle":"GPT 6.1 Sol","href":"/zh/development/comfy-router/models/openai/gpt-6-1-sol/code"},{"title":"使用 GPT 6.1 Terra 搭配 Comfy Router","sidebarTitle":"GPT 6.1 Terra","href":"/zh/development/comfy-router/models/openai/gpt-6-1-terra/code"},{"title":"将 GPT 6 Astra 与 Comfy Router 配合使用","sidebarTitle":"GPT 6 Astra","href":"/zh/development/comfy-router/models/openai/gpt-6-astra/code"},{"title":"将 GPT 6 Luna 与 Comfy Router 配合使用","sidebarTitle":"GPT 6 Luna","href":"/zh/development/comfy-router/models/openai/gpt-6-luna/code"},{"title":"将 GPT 6 Sol 与 Comfy Router 配合使用","sidebarTitle":"GPT 6 Sol","href":"/zh/development/comfy-router/models/openai/gpt-6-sol/code"},{"title":"将 GPT Image 1.5 与 Comfy Router 配合使用","sidebarTitle":"GPT Image 1.5","href":"/zh/development/comfy-router/models/openai/gpt-image-1-5/code"},{"title":"将 GPT Image 2.5 Flare 与 Comfy Router 搭配使用","sidebarTitle":"GPT Image 2.5 Flare","href":"/zh/development/comfy-router/models/openai/gpt-image-2-5-flare/code"},{"title":"通过 Comfy Router 使用 GPT Image 2.5 Sunburst","sidebarTitle":"GPT Image 2.5 Sunburst","href":"/zh/development/comfy-router/models/openai/gpt-image-2-5-sunburst/code"},{"title":"将 GPT Image 2.6 Flare 与 Comfy Router 配合使用","sidebarTitle":"GPT Image 2.6 Flare","href":"/zh/development/comfy-router/models/openai/gpt-image-2-6-flare/code"},{"title":"使用 Comfy Router 调用 GPT Image 2.6 Sunburst","sidebarTitle":"GPT Image 2.6 Sunburst","href":"/zh/development/comfy-router/models/openai/gpt-image-2-6-sunburst/code"},{"title":"将 GPT Image 2 与 Comfy Router 配合使用","sidebarTitle":"GPT Image 2","href":"/zh/development/comfy-router/models/openai/gpt-image-2/code"},{"title":"将 o3 与 Comfy Router 配合使用","sidebarTitle":"o3","href":"/zh/development/comfy-router/models/openai/o3/code"}]},{"group":"Pika","pages":[{"title":"使用 Pika 2.5 图生视频与 Comfy Router","sidebarTitle":"Pika 2.5 图生视频","href":"/zh/development/comfy-router/models/pika/pika-2-5-image-to-video/code"},{"title":"使用 Comfy Router 调用 Pika Soundtrack","sidebarTitle":"Pika Soundtrack","href":"/zh/development/comfy-router/models/pika/pika-soundtrack/code"},{"title":"通过 Comfy Router 使用 Pika Sfx","sidebarTitle":"Pika Sfx","href":"/zh/development/comfy-router/models/pika/pika-sfx/code"},{"title":"通过 Comfy Router 使用 Pika Music","sidebarTitle":"Pika Music","href":"/zh/development/comfy-router/models/pika/pika-music/code"},{"title":"使用 Pika 2.5 文生视频与 Comfy Router","sidebarTitle":"Pika 2.5 文生视频","href":"/zh/development/comfy-router/models/pika/pika-2-5-text-to-video/code"},{"title":"通过 Comfy Router 使用 Pika H3 Flex 图生视频","sidebarTitle":"Pika H3 Flex 图生视频","href":"/zh/development/comfy-router/models/pika/pika-h3-flex-image-to-video/code"},{"title":"通过 Comfy Router 使用 Pika H3 Flex 参考生视频","sidebarTitle":"Pika H3 Flex 参考生视频","href":"/zh/development/comfy-router/models/pika/pika-h3-flex-reference-to-video/code"},{"title":"使用 Comfy Router 调用 Pika H3 Flex 文生视频","sidebarTitle":"Pika H3 Flex 文生视频","href":"/zh/development/comfy-router/models/pika/pika-h3-flex-text-to-video/code"}]},{"group":"Pruna","pages":[{"title":"结合 Comfy Router 使用 P Video 2","sidebarTitle":"P Video 2","href":"/zh/development/comfy-router/models/pruna/p-video-2/code"}]},{"group":"Qwen","pages":[{"title":"搭配 Comfy Router 使用 Qwen Image 3.0 Pro","sidebarTitle":"Qwen Image 3.0 Pro","href":"/zh/development/comfy-router/models/qwen/qwen-image-3-0-pro/code"},{"title":"使用 Qwen Image 3.0 搭配 Comfy Router","sidebarTitle":"Qwen Image 3.0","href":"/zh/development/comfy-router/models/qwen/qwen-image-3-0/code"}]},{"group":"Recraft","pages":[{"title":"配合 Comfy Router 使用 Recraft V2","sidebarTitle":"Recraft V2","href":"/zh/development/comfy-router/models/recraft/recraftv2/code"},{"title":"使用 Comfy Router 调用 Recraft V3","sidebarTitle":"Recraft V3","href":"/zh/development/comfy-router/models/recraft/recraftv3/code"},{"title":"将 Recraft V4.1 Pro Vector 与 Comfy Router 配合使用","sidebarTitle":"Recraft V4.1 Pro Vector","href":"/zh/development/comfy-router/models/recraft/recraftv4-1-pro-vector/code"},{"title":"将 Recraft V4.1 Pro 与 Comfy Router 配合使用","sidebarTitle":"Recraft V4.1 Pro","href":"/zh/development/comfy-router/models/recraft/recraftv4-1-pro/code"},{"title":"通过 Comfy Router 使用 Recraft V4.1 Utility Pro Vector","sidebarTitle":"Recraft V4.1 Utility Pro Vector","href":"/zh/development/comfy-router/models/recraft/recraftv4-1-utility-pro-vector/code"},{"title":"将 Recraft V4.1 Utility Pro 与 Comfy Router 配合使用","sidebarTitle":"Recraft V4.1 Utility Pro","href":"/zh/development/comfy-router/models/recraft/recraftv4-1-utility-pro/code"},{"title":"使用 Recraft V4.1 Utility Vector 搭配 Comfy Router","sidebarTitle":"Recraft V4.1 Utility Vector","href":"/zh/development/comfy-router/models/recraft/recraftv4-1-utility-vector/code"},{"title":"将 Recraft V4.1 Utility 与 Comfy Router 搭配使用","sidebarTitle":"Recraft V4.1 Utility","href":"/zh/development/comfy-router/models/recraft/recraftv4-1-utility/code"},{"title":"结合 Comfy Router 使用 Recraft V4.1 Vector","sidebarTitle":"Recraft V4.1 Vector","href":"/zh/development/comfy-router/models/recraft/recraftv4-1-vector/code"},{"title":"搭配 Comfy Router 使用 Recraft V4.1","sidebarTitle":"Recraft V4.1","href":"/zh/development/comfy-router/models/recraft/recraftv4-1/code"},{"title":"在 Comfy Router 中使用 Recraft V4 Pro","sidebarTitle":"Recraft V4 Pro","href":"/zh/development/comfy-router/models/recraft/recraftv4-pro/code"},{"title":"将 Recraft V4 Styles Pro Vector 与 Comfy Router 配合使用","sidebarTitle":"Recraft V4 Styles Pro Vector","href":"/zh/development/comfy-router/models/recraft/recraftv4-styles-pro-vector/code"},{"title":"将 Recraft V4 Styles Pro 与 Comfy Router 配合使用","sidebarTitle":"Recraft V4 Styles Pro","href":"/zh/development/comfy-router/models/recraft/recraftv4-styles-pro/code"},{"title":"使用 Recraft V4 Styles Vector 与 Comfy Router","sidebarTitle":"Recraft V4 Styles Vector","href":"/zh/development/comfy-router/models/recraft/recraftv4-styles-vector/code"},{"title":"使用 Recraft V4 Styles 搭配 Comfy Router","sidebarTitle":"Recraft V4 Styles","href":"/zh/development/comfy-router/models/recraft/recraftv4-styles/code"},{"title":"搭配 Comfy Router 使用 Recraft V4","sidebarTitle":"Recraft V4","href":"/zh/development/comfy-router/models/recraft/recraftv4/code"}]},{"group":"Runway","pages":[{"title":"将 Aleph 2 与 Comfy Router 配合使用","sidebarTitle":"Aleph 2","href":"/zh/development/comfy-router/models/runway/aleph2/code"},{"title":"通过 Comfy Router 使用 Gen 4 Image","sidebarTitle":"Gen 4 Image","href":"/zh/development/comfy-router/models/runway/gen4-image/code"},{"title":"将 Gen 4 Turbo 与 Comfy Router 配合使用","sidebarTitle":"Gen 4 Turbo","href":"/zh/development/comfy-router/models/runway/gen4-turbo/code"}]},{"group":"Stability","pages":[{"title":"将 Control Sketch 与 Comfy Router 配合使用","sidebarTitle":"Control Sketch","href":"/zh/development/comfy-router/models/stability/control-sketch/code"},{"title":"搭配 Comfy Router 使用 Control Structure","sidebarTitle":"Control Structure","href":"/zh/development/comfy-router/models/stability/control-structure/code"},{"title":"使用 Comfy Router 调用 Control Style Transfer","sidebarTitle":"Control Style Transfer","href":"/zh/development/comfy-router/models/stability/control-style-transfer/code"},{"title":"将 Control Style 与 Comfy Router 配合使用","sidebarTitle":"Control Style","href":"/zh/development/comfy-router/models/stability/control-style/code"},{"title":"通过 Comfy Router 使用 Edit Erase","sidebarTitle":"Edit Erase","href":"/zh/development/comfy-router/models/stability/edit-erase/code"},{"title":"通过 Comfy Router 使用 Edit Inpaint","sidebarTitle":"Edit Inpaint","href":"/zh/development/comfy-router/models/stability/edit-inpaint/code"},{"title":"使用 Comfy Router 调用 Edit Outpaint","sidebarTitle":"Edit Outpaint","href":"/zh/development/comfy-router/models/stability/edit-outpaint/code"},{"title":"使用 Comfy Router 调用 Edit Remove Background","sidebarTitle":"移除背景","href":"/zh/development/comfy-router/models/stability/edit-remove-background/code"},{"title":"将 Edit Replace Background And Relight 与 Comfy Router 配合使用","sidebarTitle":"Edit Replace Background And Relight","href":"/zh/development/comfy-router/models/stability/edit-replace-background-and-relight/code"},{"title":"使用 Comfy Router 调用 Edit Search And Recolor","sidebarTitle":"Edit Search And Recolor","href":"/zh/development/comfy-router/models/stability/edit-search-and-recolor/code"},{"title":"使用 Comfy Router 调用 Edit Search And Replace","sidebarTitle":"编辑、搜索与替换","href":"/zh/development/comfy-router/models/stability/edit-search-and-replace/code"},{"title":"将 Generate Core 与 Comfy Router 搭配使用","sidebarTitle":"Generate Core","href":"/zh/development/comfy-router/models/stability/generate-core/code"},{"title":"使用 Generate Ultra 搭配 Comfy Router","sidebarTitle":"Generate Ultra","href":"/zh/development/comfy-router/models/stability/generate-ultra/code"},{"title":"通过 Comfy Router 使用 Sd3.5 Flash","sidebarTitle":"Sd3.5 Flash","href":"/zh/development/comfy-router/models/stability/sd3-5-flash/code"},{"title":"通过 Comfy Router 使用 Sd3.5 Large Turbo","sidebarTitle":"Sd3.5 Large Turbo","href":"/zh/development/comfy-router/models/stability/sd3-5-large-turbo/code"},{"title":"通过 Comfy Router 使用 Sd3.5 Large","sidebarTitle":"Sd3.5 Large","href":"/zh/development/comfy-router/models/stability/sd3-5-large/code"},{"title":"通过 Comfy Router 使用 Sd3.5 Medium","sidebarTitle":"Sd3.5 Medium","href":"/zh/development/comfy-router/models/stability/sd3-5-medium/code"},{"title":"使用 Sdxl 1.0 图生图与 Comfy Router","sidebarTitle":"Sdxl 1.0 图生图","href":"/zh/development/comfy-router/models/stability/sdxl-1-0-image-to-image/code"},{"title":"将 Sdxl 1.0 Masking 与 Comfy Router 配合使用","sidebarTitle":"Sdxl 1.0 Masking","href":"/zh/development/comfy-router/models/stability/sdxl-1-0-masking/code"},{"title":"使用 Comfy Router 调用 Sdxl 1.0 文生图","sidebarTitle":"Sdxl 1.0 文生图","href":"/zh/development/comfy-router/models/stability/sdxl-1-0-text-to-image/code"},{"title":"使用 Comfy Router 调用 Stable Fast 3D","sidebarTitle":"Stable Fast 3D","href":"/zh/development/comfy-router/models/stability/stable-fast-3d/code"},{"title":"通过 Comfy Router 使用 Stable Point Aware 3D","sidebarTitle":"Stable Point Aware 3D","href":"/zh/development/comfy-router/models/stability/stable-point-aware-3d/code"},{"title":"使用 Comfy Router 调用 Upscale Conservative","sidebarTitle":"Upscale Conservative","href":"/zh/development/comfy-router/models/stability/upscale-conservative/code"},{"title":"通过 Comfy Router 使用 Upscale Creative","sidebarTitle":"Upscale Creative","href":"/zh/development/comfy-router/models/stability/upscale-creative/code"},{"title":"使用 Upscale Fast 与 Comfy Router","sidebarTitle":"Upscale Fast","href":"/zh/development/comfy-router/models/stability/upscale-fast/code"}]},{"group":"Synclabs","pages":[{"title":"通过 Comfy Router 使用 Sync 3","sidebarTitle":"Sync 3","href":"/zh/development/comfy-router/models/synclabs/sync-3/code"}]},{"group":"Tencent","pages":[{"title":"使用混元 3D Part 配合 Comfy Router","sidebarTitle":"混元 3D Part","href":"/zh/development/comfy-router/models/tencent/hunyuan-3d-part/code"},{"title":"通过 Comfy Router 使用混元3D智能拓扑","sidebarTitle":"混元3D智能拓扑","href":"/zh/development/comfy-router/models/tencent/hunyuan-3d-smart-topology/code"},{"title":"使用 Comfy Router 调用混元3D 纹理编辑","sidebarTitle":"混元 3D Texture Edit","href":"/zh/development/comfy-router/models/tencent/hunyuan-3d-texture-edit/code"},{"title":"通过 Comfy Router 使用混元 3D Uv","sidebarTitle":"混元 3D Uv","href":"/zh/development/comfy-router/models/tencent/hunyuan-3d-uv/code"}]},{"group":"Veed","pages":[{"title":"通过 Comfy Router 使用 Clean Audio Video","sidebarTitle":"Clean Audio Video","href":"/zh/development/comfy-router/models/veed/clean-audio-video/code"},{"title":"通过 Comfy Router 使用 Clean Audio","sidebarTitle":"Clean Audio","href":"/zh/development/comfy-router/models/veed/clean-audio/code"},{"title":"使用 Fabric 1.0 与 Comfy Router","sidebarTitle":"Fabric 1.0","href":"/zh/development/comfy-router/models/veed/fabric-1-0/code"},{"title":"使用 Comfy Router 调用 Lipsync 2.0","sidebarTitle":"Lipsync 2.0","href":"/zh/development/comfy-router/models/veed/lipsync-2-0/code"},{"title":"使用 Comfy Router 进行快速视频背景移除","sidebarTitle":"快速视频背景移除","href":"/zh/development/comfy-router/models/veed/video-background-removal-fast/code"},{"title":"使用 Comfy Router 进行视频背景移除绿幕处理","sidebarTitle":"视频背景移除绿幕","href":"/zh/development/comfy-router/models/veed/video-background-removal-green-screen/code"},{"title":"通过 Comfy Router 使用视频背景移除","sidebarTitle":"视频背景移除","href":"/zh/development/comfy-router/models/veed/video-background-removal/code"}]},{"group":"Veo","pages":[{"title":"将 Veo 3.1 Fast Generate 001 与 Comfy Router 配合使用","sidebarTitle":"Veo 3.1 Fast Generate 001","href":"/zh/development/comfy-router/models/veo/veo-3-1-fast-generate-001/code"},{"title":"使用 Comfy Router 调用 Veo 3.1 Generate 001","sidebarTitle":"Veo 3.1 Generate 001","href":"/zh/development/comfy-router/models/veo/veo-3-1-generate-001/code"},{"title":"使用 Comfy Router 调用 Veo 3.1 Lite Generate 001","sidebarTitle":"Veo 3.1 Lite Generate 001","href":"/zh/development/comfy-router/models/veo/veo-3-1-lite-generate-001/code"}]},{"group":"Wan","pages":[{"title":"通过 Comfy Router 使用 HappyHorse 1.0 I2V","sidebarTitle":"HappyHorse 1.0 I2V","href":"/zh/development/comfy-router/models/wan/happyhorse-1-0-i2v/code"},{"title":"使用 HappyHorse 1.0 R2V 搭配 Comfy Router","sidebarTitle":"HappyHorse 1.0 R2V","href":"/zh/development/comfy-router/models/wan/happyhorse-1-0-r2v/code"},{"title":"使用 Comfy Router 调用 HappyHorse 1.0 T2V","sidebarTitle":"HappyHorse 1.0 T2V","href":"/zh/development/comfy-router/models/wan/happyhorse-1-0-t2v/code"},{"title":"使用 Comfy Router 调用 HappyHorse 1.0 Video Edit","sidebarTitle":"HappyHorse 1.0 视频编辑","href":"/zh/development/comfy-router/models/wan/happyhorse-1-0-video-edit/code"},{"title":"将 HappyHorse 1.1 I2V 与 Comfy Router 配合使用","sidebarTitle":"HappyHorse 1.1 I2V","href":"/zh/development/comfy-router/models/wan/happyhorse-1-1-i2v/code"},{"title":"通过 Comfy Router 使用 HappyHorse 1.1 R2V","sidebarTitle":"HappyHorse 1.1 R2V","href":"/zh/development/comfy-router/models/wan/happyhorse-1-1-r2v/code"},{"title":"使用 Comfy Router 调用 HappyHorse 1.1 T2V","sidebarTitle":"HappyHorse 1.1 T2V","href":"/zh/development/comfy-router/models/wan/happyhorse-1-1-t2v/code"},{"title":"通过 Comfy Router 使用 Wan 2.5 I2I Preview","sidebarTitle":"Wan 2.5 I2I Preview","href":"/zh/development/comfy-router/models/wan/wan2-5-i2i-preview/code"},{"title":"使用 Comfy Router 调用 Wan 2.5 I2V Preview","sidebarTitle":"Wan 2.5 I2V Preview","href":"/zh/development/comfy-router/models/wan/wan2-5-i2v-preview/code"},{"title":"将 Wan 2.5 T2I Preview 与 Comfy Router 配合使用","sidebarTitle":"Wan 2.5 T2I Preview","href":"/zh/development/comfy-router/models/wan/wan2-5-t2i-preview/code"},{"title":"使用 Comfy Router 调用 Wan 2.5 T2V Preview","sidebarTitle":"Wan 2.5 T2V Preview","href":"/zh/development/comfy-router/models/wan/wan2-5-t2v-preview/code"},{"title":"使用 Comfy Router 调用 Wan 2.6 I2V","sidebarTitle":"Wan 2.6 I2V","href":"/zh/development/comfy-router/models/wan/wan2-6-i2v/code"},{"title":"通过 Comfy Router 使用 Wan 2.6 R2V","sidebarTitle":"Wan 2.6 R2V","href":"/zh/development/comfy-router/models/wan/wan2-6-r2v/code"},{"title":"通过 Comfy Router 使用 Wan 2.6 T2V","sidebarTitle":"Wan 2.6 T2V","href":"/zh/development/comfy-router/models/wan/wan2-6-t2v/code"},{"title":"使用 Comfy Router 调用 Wan 2.7 I2V","sidebarTitle":"Wan 2.7 I2V","href":"/zh/development/comfy-router/models/wan/wan2-7-i2v/code"},{"title":"使用 Comfy Router 调用 Wan 2.7 R2V","sidebarTitle":"Wan 2.7 R2V","href":"/zh/development/comfy-router/models/wan/wan2-7-r2v/code"},{"title":"使用 Comfy Router 调用 Wan 2.7 T2V","sidebarTitle":"Wan 2.7 T2V","href":"/zh/development/comfy-router/models/wan/wan2-7-t2v/code"},{"title":"搭配 Comfy Router 使用 Wan 2.7 Video Edit","sidebarTitle":"Wan 2.7 Video Edit","href":"/zh/development/comfy-router/models/wan/wan2-7-videoedit/code"},{"title":"将 Wan 3.0 Video Prime 与 Comfy Router 配合使用","sidebarTitle":"Wan 3.0 Video Prime","href":"/zh/development/comfy-router/models/wan/wan3-0-video-prime/code"},{"title":"通过 Comfy Router 使用 Wan 3.0 Video","sidebarTitle":"Wan 3.0 Video","href":"/zh/development/comfy-router/models/wan/wan3-0-video/code"}]},{"group":"WaveSpeed","pages":[{"title":"将 Flashvsr 与 Comfy Router 配合使用","sidebarTitle":"Flashvsr","href":"/zh/development/comfy-router/models/wavespeed/flashvsr/code"},{"title":"将 Seedvr 2 与 Comfy Router 配合使用","sidebarTitle":"Seedvr 2","href":"/zh/development/comfy-router/models/wavespeed/seedvr2/code"},{"title":"将 Ultimate Image Upscaler 与 Comfy Router 配合使用","sidebarTitle":"Ultimate Image Upscaler","href":"/zh/development/comfy-router/models/wavespeed/ultimate-image-upscaler/code"}]},{"group":"xAI","pages":[{"title":"将 Grok Imagine Image 2.0 与 Comfy Router 配合使用","sidebarTitle":"Grok Imagine Image 2.0","href":"/zh/development/comfy-router/models/xai/grok-imagine-image-2-0/code"},{"title":"配合 Comfy Router 使用 Grok Imagine Image Pro","sidebarTitle":"Grok Imagine Image Pro","href":"/zh/development/comfy-router/models/xai/grok-imagine-image-pro/code"},{"title":"配合 Comfy Router 使用 Grok Imagine Image Quality","sidebarTitle":"Grok Imagine Image Quality","href":"/zh/development/comfy-router/models/xai/grok-imagine-image-quality/code"},{"title":"使用 Grok Imagine Image 与 Comfy Router","sidebarTitle":"Grok Imagine Image","href":"/zh/development/comfy-router/models/xai/grok-imagine-image/code"},{"title":"将 Grok Imagine Video 1.5 Lite 与 Comfy Router 配合使用","sidebarTitle":"Grok Imagine Video 1.5 Lite","href":"/zh/development/comfy-router/models/xai/grok-imagine-video-1-5-lite/code"},{"title":"结合 Comfy Router 使用 Grok Imagine Video 1.5 Preview","sidebarTitle":"Grok Imagine Video 1.5 Preview","href":"/zh/development/comfy-router/models/xai/grok-imagine-video-1-5-preview/code"},{"title":"将 Grok Imagine Video 1.5 与 Comfy Router 配合使用","sidebarTitle":"Grok Imagine Video 1.5","href":"/zh/development/comfy-router/models/xai/grok-imagine-video-1-5/code"},{"title":"搭配 Comfy Router 使用 Grok Imagine Video","sidebarTitle":"Grok Imagine Video","href":"/zh/development/comfy-router/models/xai/grok-imagine-video/code"}]},{"group":"Openrouter","pages":[{"title":"配合 Comfy Router 使用 Chat Completions","sidebarTitle":"Chat Completions","href":"/zh/development/comfy-router/models/openrouter/chat-completions/code"}]}]}]},{"group":"Comfy CLI","pages":[{"title":"Comfy CLI 入门","sidebarTitle":"快速开始","href":"/zh/comfy-cli/getting-started"},{"title":"Comfy CLI 命令参考","sidebarTitle":"命令参考","href":"/zh/comfy-cli/reference"},{"title":"Comfy CLI 故障排查:安装与环境修复","sidebarTitle":"故障排除","href":"/zh/comfy-cli/troubleshooting"}]},{"group":"规范","pages":[{"group":"Workflow JSON","pages":[{"title":"工作流 JSON","href":"/zh/specs/workflow_json"},{"title":"工作流 JSON 0.4","href":"/zh/specs/workflow_json_0.4"}]},{"group":"节点定义","pages":[{"title":"节点定义 JSON","href":"/zh/specs/nodedef_json"},{"title":"节点定义 JSON 1.0","href":"/zh/specs/nodedef_json_1_0"}]}]},{"group":"Demo Apps","pages":[{"title":"演示应用","mode":"wide","href":"/zh/development/samples/overview"}]}]},{"tab":"自定义节点","pages":[{"title":"自定义节点","href":"/zh/custom-nodes/intro"}]},{"tab":"支持","pages":[{"title":"支持","href":"/zh/support/contact-support"}]},{"tab":"Cloud API 参考文档","pages":[{"title":"Cloud API 参考文档","href":"/zh/api-reference/cloud/workflow/get-information-about-current-prompt-executions"}]}]},{"language":"ja","footer":{"socials":{"github":"https://github.com/Comfy-Org/ComfyUI","x":"https://x.com/ComfyUI","discord":"https://discord.com/invite/comfyorg","youtube":"https://www.youtube.com/@comfyorg"},"links":[{"header":"リソース","items":[{"label":"インストール","href":"/ja/installation/system_requirements"},{"label":"チュートリアル","href":"/ja/tutorials/basic/text-to-image"},{"label":"開発","href":"/ja/development/overview"}]},{"header":"プロダクト","items":[{"label":"機能","href":"https://www.comfy.org/?utm_source=docs"},{"label":"ギャラリー","href":"https://www.comfy.org/gallery?utm_source=docs"},{"label":"ダウンロード","href":"https://www.comfy.org/download?utm_source=docs"}]},{"header":"会社情報","items":[{"label":"概要","href":"https://www.comfy.org/about?utm_source=docs"},{"label":"採用情報","href":"https://www.comfy.org/careers?utm_source=docs"},{"label":"利用規約","href":"https://www.comfy.org/terms-of-service?utm_source=docs"},{"label":"プライバシーポリシー","href":"https://www.comfy.org/privacy-policy?utm_source=docs"}]}]},"navbar":{"links":[{"href":"/ja/tutorials/partner-nodes/pricing","label":"料金"},{"href":"https://comfy.org/download?utm_source=docs","label":"ダウンロード"}],"primary":{"type":"button","label":"Comfy Cloud","href":"https://comfy.org/cloud?utm_source=docs"}},"pages":[{"title":"ja","href":"/ja/api-reference/router/comfy-router/list-the-models-comfy-router-can-run"}]},{"language":"ko","footer":{"socials":{"github":"https://github.com/Comfy-Org/ComfyUI","x":"https://x.com/ComfyUI","discord":"https://discord.com/invite/comfyorg","youtube":"https://www.youtube.com/@comfyorg"},"links":[{"header":"리소스","items":[{"label":"설치","href":"/ko/installation/system_requirements"},{"label":"튜토리얼","href":"/ko/tutorials/basic/text-to-image"},{"label":"개발","href":"/ko/development/overview"}]},{"header":"제품","items":[{"label":"기능","href":"https://www.comfy.org/?utm_source=docs"},{"label":"갤러리","href":"https://www.comfy.org/gallery?utm_source=docs"},{"label":"다운로드","href":"https://www.comfy.org/download?utm_source=docs"}]},{"header":"회사 정보","items":[{"label":"소개","href":"https://www.comfy.org/about?utm_source=docs"},{"label":"채용","href":"https://www.comfy.org/careers?utm_source=docs"},{"label":"이용약관","href":"https://www.comfy.org/terms-of-service?utm_source=docs"},{"label":"개인정보 처리방침","href":"https://www.comfy.org/privacy-policy?utm_source=docs"}]}]},"navbar":{"links":[{"href":"/ko/tutorials/partner-nodes/pricing","label":"가격 정책"},{"href":"https://comfy.org/download?utm_source=docs","label":"다운로드"}],"primary":{"type":"button","label":"Comfy Cloud","href":"https://comfy.org/cloud?utm_source=docs"}},"pages":[{"title":"ko","href":"/ko/api-reference/router/comfy-router/list-the-models-comfy-router-can-run"}]},{"language":"es","pages":[{"title":"es","href":"/es/api-reference/router/comfy-router/list-the-models-comfy-router-can-run"}]}]},"basePath":"","title":"List the models Comfy Router can run. - ComfyUI","buildId":"dpl_42LwxxU1kGN114H4XzwGYXa9ZDah","deploymentId":"dpl_42LwxxU1kGN114H4XzwGYXa9ZDah","dynamics":["c2aa737661","2b790eda92","b0f84f53ff","a9bb46feb3"],"chunks":["OpenApiReferenceProvider-C-hEh3I2.js","utils-CnF4QiUD.js","cloneDeep-izs3Hzxb.js","_Uint8Array-BVIqJ90l.js","isObjectLike-C8vmzPEN.js","isArray-Dcu1dQ5t.js","utils-C7PVNbel.js","toString-BX-iyh8G.js","isSymbol-CSH9U2Gm.js","addUniqueKeysToSchema-8ytevL2A.js","snippetPresets-BT1alx3v.js","style-to-object-BY3JEAdk.js","v4-DIgz8zy6.js","conditional-requirements-CFM0id0Q.js","ApiExamples-BLXC0i_W.js","isEqual-BfaiBbcl.js","usePlaygroundInputsStore-CyLIGI9_.js","shallow-DC0lGVIW.js","getApiPlaygroundDisplay-CgDQ85tN.js","CodeSnippet-6-I-Gpv9.js","CodeSnippets-BAk1kDdu.js","Page-C945cI-W.js","getPastedJson-B276RuWN.js","icon-BqP8etTm.js","MethodPill-DM6sbxG1.js","DuplicateIcon-PyJglOaE.js","createIcon-CMeOMAc_.js","DynamicMarkdown-BoyYAA8K.js","SectionHeading-DAYv3gvZ.js","axios-vUo_bRtp.js","Markdown-BhTTmazV.js","clsx-CR-RHrVC.js","lib-DuNMoZ05.js","lib-B8HbPgcC.js","html-void-elements-mCSPHezU.js","dist-CBUH0-zX.js","katex-YMM-FMeD.js","marked.esm-CZWythHk.js","AssistantPageLink-BmARFYNC.js"]}