itsjwill/vanta ? reverse-engineered prompt

Build me a local AI video creator that feels like a free alternative to Synthesia, Runway, CapCut, Descript, and ElevenLabs. I want to be able to drop in a script, clone a voice from a short sample, make a talking head avatar from a photo, generate animated captions, add background removal, music, transitions, and simple motion graphics, then export a finished video.

It should open in the browser with a visual editor and timeline, and render videos through Remotion. Keep it all running locally if possible, with no API keys or subscriptions. Make it easy to swap between voiceover, avatar video, text to video, and regular editing tools. If you need to check current docs for any of the AI models or Remotion, go ahead and look them up online.

I’d like a clean starter experience, with a demo project I can render right away and enough structure that I can keep adding new video effects later.