Vidu's Q1 has dropped its Reference-to-Image tool in the middle of 2025 that takes up to 7 pictures to make a new one, keeping characters and stuff looking the same across all of them.
You use it through Vidu’s app or APIs. There's no info yet on local model files or VRAM needs.
It’s a tool that lets you give it a few images like a person, outfit, and background, plus a prompt. It uses all that to make one new picture that holds onto the same look and style. Handy for things like brand sheets, product shots, or storyboards.
If you'd like to access this model, you can explore the following possibilities:
