ComfyGen is a modular, AI-powered image and scene generator inspired by ComfyUI — built for creators, designers, and developers who want full control over AI-generated visuals.
Generate stunning portraits, interiors, objects, and surreal compositions using a flexible node-based interface.
- Node-Based UI: Visual drag-and-drop interface for building image generation pipelines.
- Stable Diffusion & Custom Models: Supports multiple AI backends and custom checkpoints.
- Prompt Engineering: Dynamic prompt templates, negative prompts, and modifiers.
- Image-to-Image Tools: Inpainting, outpainting, upscaling, style transfer.
- Multi-resolution Output: From thumbnails to 4K renders.
- Batch Generation: Create hundreds of variations from one prompt.
- Plugin System: Extend functionality with community or custom nodes.
- Interior and architectural renders.
- Stylized concept art and illustrations.
- Product mockups.
- AI avatars and portraits.
- Social media visuals.
- Import custom .ckpt, .safetensors, or .pt diffusion models.
- Extend nodes with custom logic using Python.
- Compatible with ComfyUI workflows (via import/export).
ups.mp4
- All generated images are saved in /outputs/.
- Includes prompt metadata and seed values.
- Optional autosave and versioning enabled.
- OS: Windows 10/11, macOS 12+, or Linux.
- GPU: NVIDIA (6GB+ VRAM) recommended.
- Python: 3.10+.
- Storage: ~10GB for models and outputs.
Comfy is released under the MIT License — free to use, modify, and distribute.