Shital InfotechPrompt-to-Reel — AI video generation platform
Type a prompt, get a finished narrated video — script, voiceover, avatar, b-roll and a rendered edit.
A project-based studio where a user describes a video, refines the script through chat, and the platform builds it end to end — LLM script, TTS voiceover, talking avatar, generative b-roll, all composited by FFmpeg.
The problem
Producing a narrated video normally means writing a script, recording voiceover, filming or animating visuals, and editing it all together — separate tools and separate skills for each step.
How we built it
A chat-driven scripting flow: describe the video, refine the script by chatting with the AI, then lock it in. Once finalized, the platform breaks it into scenes, generates the voiceover, avatar footage and b-roll for each one, and composites everything into a finished video automatically.
What it does
- Draft and refine the script conversationally before locking it in — no video editing skill required
- Choose from a library of talking avatars, including custom/private ones, with support for two avatars in one scene
- Voice and captions stay perfectly synced to the visuals, scene by scene
- Every project is saved to a private workspace with render status you can check in on
- Renders happen in the background — start a render and come back to a finished video