Skip to main content
Ideogram 4.0 is the latest text-to-image model from Ideogram, released as an open-source model that runs entirely on your own hardware. It delivers superior photorealistic quality, accurate text rendering, and precise style control. You can use either plain natural language or structured JSON prompts for fine-grained control over layout, colors, and in-image text.
Make sure your ComfyUI is updated.Workflows in this guide can be found in the Workflow Templates. If you can’t find them in the template, your ComfyUI may be outdated. (Desktop version’s update will delay sometime)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

​
Ideogram 4.0 Text-to-Image Workflow

Run in Comfy Cloud

Open in Comfy Cloud

Download Workflow

Download JSON or search “Ideogram v4: Text to Image” in Template Library
Ideogram 4.0 Example Output Example output from the Ideogram 4.0 model

​
Prompt Format

The open-source workflow supports two prompt modes:
  1. Natural language — Quick and easy, best for simple ideas
  2. Structured JSON — For precise control over layout, colors, and style
The workflow includes a prompt construction template that you can use with any LLM to generate matching JSON prompts. A note inside the workflow explains:
The model is trained on structured JSON captions (scene summary, style block, background, and optional per-object descriptions with bounding boxes and hex color palettes). Official inference validates prompts against that schema.

​
Ideogram 4.0 model downloads

You can find all repackaged model files at Comfy-Org/Ideogram-4 on Hugging Face.

ideogram4_fp8_scaled.safetensors

Diffusion model for Ideogram 4.0 (~13.8 GB). Place in models/diffusion_models/

ideogram4_unconditional_fp8_scaled.safetensors

Unconditional diffusion model for Ideogram 4.0 (~13.8 GB). Place in models/diffusion_models/

qwen3vl_8b_fp8_scaled.safetensors

Text encoder for Ideogram 4.0 (~8 GB). Place in models/text_encoders/

gemma4_e4b_it_fp8_scaled.safetensors

Text encoder for Ideogram 4.0 (~2 GB). Place in models/text_encoders/

flux2-vae.safetensors

VAE for Ideogram 4.0 (~335 MB). Place in models/vae/
Model storage location 
📂 ComfyUI/
├── 📂 models/
│   ├── 📂 diffusion_models/
│   │   ├── ideogram4_fp8_scaled.safetensors
│   │   └── ideogram4_unconditional_fp8_scaled.safetensors
│   ├── 📂 text_encoders/
│   │   ├── qwen3vl_8b_fp8_scaled.safetensors
│   │   └── gemma4_e4b_it_fp8_scaled.safetensors
│   └── 📂 vae/
│       └── flux2-vae.safetensors

Learn about Subgraph

This workflow uses Subgraph nodes for modular processing. Check the Subgraph documentation to learn how to customize and extend.

​
Workflow Steps

  1. Download and place the models in the correct directories (see table above)
  2. Download and drag the workflow file into ComfyUI
  3. Enter your prompt (natural language or structured JSON) in the Ideogram4 subgraph node
  4. (Optional) Adjust resolution using the ResolutionSelector node
  5. Click Run or use shortcut Ctrl(cmd) + Enter to generate the image
  6. View the result in the Save Image node — output files are saved to ComfyUI/output/

​
Note on Safety Filter

If you see “Image blocked by safety filter”, this is due to safety training baked into the model itself. ComfyUI does not add any safety filters. Try adjusting your prompt to avoid triggering the built-in safety mechanisms.