What AI model do you use?
Your personal model is a FLUX.1-dev LoRA. Edits use FLUX.1 Kontext, zoom out FLUX Fill, videos Seedance, upscaling Real-ESRGAN and try-on IDM-VTON.
We use a combination of leading open and commercial AI models, each picked for a specific job.
Your personal model
Your photos are used to train a private AI model, a LoRA on top of FLUX.1-dev. Training runs on Replicate using the ostris/flux-dev-lora-trainer and takes about 15-25 minutes.
Other tools
- Photo editing and relighting: FLUX.1 Kontext, which changes an image from a text instruction.
- Combine photos: a multi-image FLUX.1 Kontext model.
- Zoom out: FLUX Fill, which extends an image beyond its original borders (outpainting).
- AI videos: ByteDance Seedance, which turns a photo into a 5-second video.
- Upscaling: Real-ESRGAN, which increases resolution and sharpness.
- Try on clothes: IDM-VTON, a virtual try-on model.
Your model is private to you and is never used to generate images for other users.