vovuhuydeveloper/agent-content-kit ? reverse-engineered prompt

Build me an open source app that turns any URL, YouTube link, PDF, DOCX, or plain text into a short form video and can publish it automatically.

I want a simple dashboard where I can paste a source, see the pipeline progress, review the generated script and voiceover, then approve or reject the final video in Telegram or in the app. After approval, it should upload to YouTube, TikTok, and Facebook using browser automation, and it should also support scheduling new videos on a cron style calendar so it can run every day without me touching it.

Please make it work with a pluggable AI provider so I can switch between OpenAI, Claude, or Gemini, and use stock video footage plus voice generation for the final edit. It should be easy to run with Docker, but also work locally if I want. If you need to look up current docs online for any of the integrations, go ahead.