CLAP!
Loading...
Preparing your workspace
Loading...
Preparing your workspace
Promvie makes AI video in which the characters actually speak. Dialogue is written in the screenplay, assigned to a character, and rendered by the video model as spoken, lip-synced lines inside the same take as the picture, with ambience and effects. Two characters can hold a conversation in one shot, listeners react instead of freezing, and each character keeps a recognisable voice from scene to scene. There is no separate text-to-speech or lip-sync step to manage.
The script is the source. Each line has a speaker, and Promvie's shot writer paces the scene around the lines: one line per shot, room for a reaction, no more words than a person can say in the time. The scene is then rendered by a video model with native audio, such as ByteDance Seedance 2.5, MiniMax H3 Max or Wan 3.0 Prime, which generates the voice, the lip movement and the sound together.
Each character gets a short voice sample when the cast is created, and that sample travels with the character into every scene they speak in, so a gravelly detective sounds like the same gravelly detective in scene one and scene nine. Longer scenes continue from the previous part's last seconds of picture and audio, so a conversation can run past a single take without a jump.
Dialogue is performed in the language of your script. English is the most reliable; Spanish, French, German, Portuguese, Italian, Japanese and Korean work well. A take runs up to 30 seconds, which is three to five exchanges, and longer scenes continue in parts. Names and unusual words are sometimes mispronounced; rewriting the line or re-rendering that part fixes it, and re-renders cost credits like any video.
Type one idea. Promvie writes the script, casts the characters and films the scenes. No experience needed.
No credit card needed · Free script, cast and storyboards + a video preview