InstantX Team

InstantX is an independent research team that open-sources generative models out of common interests.

Research Projects

• InstantCharacter: Personalize Any Characters with a Scalable Diffusion Transformer Framework
• Regional-Prompting-RAG-FLUX: Region-Aware Text-to-Image Generation via Hard Binding and Soft Refinement
• Regional-Prompting-Attention-FLUX: Training-free Regional Prompting for Diffusion Transformers
• InstantIR: Blind Image Restoration with Instant Generative Reference
• CSGO: Content-Style Composition in Text-to-Image Generation
• InstantStyle-Plus: Style Transfer with Content-Preserving in Text-to-Image Generation
• InstantStyle: Free Lunch towards Style-Preserving in Text-to-Image Generation
• InstantID: Zero-shot Identity-Preserving Generation in Seconds

Community Projects

• Qwen-Image: Qwen-Image-ControlNet-Union, Qwen-Image-ControlNet-Inpainting
• FLUX: FLUX.1-dev-Controlnet-Canny, FLUX.1-dev-Controlnet-Union, FLUX.1-dev-IP-Adapter
• SD35: SD3.5-Large-IP-Adapter
• SD3: SD3-Controlnet-Pose, SD3-Controlnet-Canny, SD3-Controlnet-Tile

For Collaboration

For research collaboration, commercial use of our models, and open-sourcing sponsorship, please contact us via [email protected]. You can also buy us a coffee via ko-fi.

Our Sponsors

We are sponsored by Huggingface, fal.ai.