> ## Documentation Index
> Fetch the complete documentation index at: https://docs.comfy.org/llms.txt
> Use this file to discover all available pages before exploring further.
# FLUX 3 Max in ComfyUI: Keyframed Video with Native Audio
> Generate 5 to 20 second videos in ComfyUI with FLUX 3 Max: text to video, keyframe-controlled image to video with up to 10 images, and native audio.
**FLUX 3 Max** is Black Forest Labs' video model on the FLUX 3 foundation, served through fal and available in ComfyUI as Partner Nodes. It generates a clip from a text prompt, or animates a set of still images into one continuous shot.
The difference from the standard FLUX 3 video nodes is control over the timeline. Image to video takes 1 to 10 keyframe images: one image opens the clip, two become its first and last frame, and more are either spread evenly across the clip or pinned to the seconds you choose. Every run also produces synchronized audio, including ambience, speech, and sound effects.
Both nodes are cloud API nodes. They need a Comfy account with credits, download no model, and bill per second of output; see [Partner Node pricing](/tutorials/partner-nodes/pricing) for the rates.
## What FLUX 3 Max is good at
* **Keyframe control**: place up to 10 images on the timeline and let the model fill in the motion between them
* **Native audio**: every clip ships with synchronized ambience, speech, and effects, or you can turn the audio track off
* **Clips from 5 to 20 seconds**: one generation covers a full shot instead of stitched segments
* **Prompt expansion**: `reasoning_level` decides how much the model rewrites and plans the prompt before generating, at the same price either way
* **Flexible framing**: 720p or 1080p at 24 fps, from 21:9 through 9:16, or `auto` to follow the input image
## Example outputs
Text to video from a single prompt, with synchronized audio:
A three-shot sequence from one keyframe image, with hard cuts written into the prompt:
> [!TIP]
>
> To use the Partner Nodes, you need to ensure that you are logged in properly and using a permitted network environment. Please refer to the [Partner Nodes Overview](/tutorials/partner-nodes/overview) section of the documentation to understand the specific requirements for using the Partner Nodes.
> [!TIP]
>
> **Local users**
>
> Make sure your ComfyUI is updated.
>
> * [Download ComfyUI](https://www.comfy.org/download)
> * [Update Guide](/installation/update_comfyui)
>
> Workflows in this guide can be found in the [Workflow Templates](/interface/features/template).
> If you can't find them in the template, your ComfyUI may be outdated.
>
> If nodes are missing when loading a workflow, possible reasons:
>
> 1. You are not using the latest ComfyUI version (Nightly version)
> 2. Some nodes failed to import at startup
>
> **Cloud users**
>
> * [Cloud](https://cloud.comfy.org) will update after ComfyUI stable release.
>
> So, if you find any core node missing in this document, it might be because the new core nodes have not yet been released in the latest stable version. Please wait for the next stable release.
## FLUX 3 Max text-to-video workflow
Describe the shot, choose a length, and run. Sound, motion, and camera behavior are generated together from the same prompt, so describe the audio you want in the same paragraph as the action.
> **[Download Workflow](https://github.com/Comfy-Org/workflow_templates/blob/main/templates/api_flux3_max_t2v.json)**
>
> Download JSON or search "FLUX 3 Max: Text to Video" in Template Library
**How to use**
1. **Write the prompt**: describe the scene, then the sound. Separate ambience, music, and speech so the audio layers do not compete.
2. **Set the clip**: `duration` runs from 5 to 20 seconds, `resolution` is `720p` or `1080p`, and `aspect_ratio` defaults to `auto`.
3. **Leave `reasoning_level` on `high`** for complex prompts with cuts or several subjects. It takes longer and costs the same as `low`.
4. **Run** and save the result with `SaveVideo`.
## FLUX 3 Max image-to-video workflow
Turn stills into one moving shot. With a single image the clip opens on it and the model invents the rest of the action; with several, the images anchor the composition at points along the timeline.
> **[Download Workflow](https://github.com/Comfy-Org/workflow_templates/blob/main/templates/api_flux3_max_i2v.json)**
>
> Download JSON or search "FLUX 3 Max: Image to Video" in Template Library
**Input material**
> **[hovering\_spaceship.png](https://raw.githubusercontent.com/Comfy-Org/workflow_templates/main/input/hovering_spaceship.png)**
>
> Load into the first keyframe slot of the FLUX 3 Max Image to Video node
**How to use**
1. **Add the images**: connect 1 to 10 images to the `keyframes` inputs, in playback order.
2. **Choose the placement**: `spread across the clip` puts the first image at the start, the last at the end, and the rest evenly between. `at times` lets you pin each image to a second you type, comma-separated and increasing, such as `0, 2.5, 5`.
3. **Describe the motion and the sound**: the prompt tells the model how the scene should move and what it should sound like. Write cuts explicitly, as `SHOT ONE`, `HARD CUT`, `SHOT TWO`.
4. **Set the clip**: `duration`, `resolution`, and `aspect_ratio` behave as in the text-to-video node.
5. **Run** and save the result with `SaveVideo`.
## Node controls
Both nodes share these inputs:
| Control | Values | Effect |
| - | - | - |
| `prompt` | up to 50,000 characters | The shot, the camera behavior, and the audio |
| `aspect_ratio` | `auto`, `21:9`, `2:1`, `16:9`, `4:3`, `1:1`, `3:4`, `9:16` | `auto` is 16:9 for text prompts and follows the first image otherwise |
| `duration` | 5 to 20 seconds | Clip length |
| `resolution` | `720p`, `1080p` | Output resolution at 24 fps |
| `reasoning_level` | `low`, `high` | How much the model expands the prompt before generating; the price does not change |
| `generate_audio` | on by default | Turn it off for a silent video |
| `seed` | default 42 | Changing it re-runs the node; results are non-deterministic regardless of the seed |
Image to video adds `keyframes` (1 to 10 images) and `placement` (`spread across the clip` or `at times`).
## Limits
* A clip is 5 to 20 seconds at 24 fps. There is no continuation input on these nodes, so longer pieces are separate generations.
* At most 10 keyframe images, and at least one.
* With `at times`, give exactly one time per image, in increasing order, and keep the last time inside the clip.
* Audio is generated with the video and cannot be edited separately before it comes back.
* The seed does not make runs reproducible. Re-running the same graph gives a different result.
## Get started
1. Update ComfyUI to the latest version
2. Go to Template Library, search for `FLUX 3 Max`, and open either template
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.