Vellux trains a private likeness model for each of your characters, then generates on-brand content in one canvas — text-to-image, inpainting, and an AI agent that handles the busywork.
Start generating in minutes.
By creating an account you confirm you are 18 or older and agree to the Terms & Privacy Policy.
Four connected tools that keep every character on-model, from first train to final export.
Train a private likeness LoRA per character — one per model. Lock in a consistent face, body and style that shows up identically in every generation.
Text-to-image and inpainting in a single canvas. Prompt a scene, mask a region, and regenerate just the parts you want — no round-trips.
An assistant that browses your references, writes prompts, and runs multi-step workflows for you — from a single instruction to a finished set.
Every generation organized, tagged and searchable. Reuse a prompt, remix an asset, or attach it as a reference in a single click.
Upload a few reference photos. Vellux trains a private likeness model — no prompt engineering required.
Describe the shot in Studio, or let the Agent do it. Every result stays true to your character.
Export in any format, batch a full set, and keep your library organized in the Gallery.
No per-image fees. Your models and content stay private and portable.
Train your first character free. Generate on-brand content in minutes — you own every model and every asset.