Photo
Identity
A scene from scratch from a prompt while keeping your face — no required interior reference.
Identity generation: you describe the scene in text, the model builds the frame and tries to keep the face from your photo.
You need: a face photo and a text prompt. A scene reference isn’t required.
Stability is slightly lower than classic faceswap into a ready frame.
Попробовать Identity
Запуск генерации — в Telegram-боте. На сайте пока только описание.