How raunakpatil/youtube-agentic-ai-studio works
This is a faceless YouTube video production studio. Its principal workflow researches a channel-fit topic, writes a long-form or Shorts script, gathers stock imagery, generates narration, renders a video, pauses for human approval, and uploads approved work. The CLI pipeline explicitly orchestrates these stages; the GUI is a separate documented control surface. The README also describes model fallback and background music. Some GUI and renderer internals were not fully sampled, so their specific connections are shown only where documentation supports them.
Main entry points: pipeline.py CLI pipeline · gui.py Web GUI · creator.py Video renderer · app.py Review dashboard
Components 18
Studio control
- CLI pipeline
pipeline.py - Web GUI
gui.py - Studio settings
config.py
Topic and script
- Topic research
agents/researcher.py - Gemini fallback client
agents/gemini_client.py - Script writing
agents/scriptwriter.py
Media production
- Stock images
video/stock.py - Pexels
- Voice narration
video/narrator.py - Edge TTS
- Video renderer
video/creator.py - Background music
video/music.py - Rendered video
Review and publishing
- Review dashboard
review/app.py - YouTube uploader
uploader/youtube.py - YouTube
Other components
- Studio user
- Human reviewer
Connections 20
- Studio user → CLI pipeline: runs pipeline
README.md - Studio user → Web GUI: uses studio
README.md - CLI pipeline → Topic research: calls
pipeline.py - Topic research → Gemini fallback client: generates with
agents/researcher.py - CLI pipeline → Script writing: calls
pipeline.py - Script writing → Gemini fallback client: generates with
agents/scriptwriter.py - CLI pipeline → Stock images: downloads images
pipeline.py - Stock images → Pexels: searches and downloads
video/stock.py - CLI pipeline → Voice narration: generates narration
pipeline.py - Voice narration → Edge TTS: synthesizes speech
README.md - CLI pipeline → Video renderer: renders video
pipeline.py - Video renderer → Background music: mixes background music
README.md - Video renderer → Rendered video: produces
- CLI pipeline → Review dashboard: starts review
pipeline.py - Review dashboard → Human reviewer: presents video for decision
review/app.py - Review dashboard → Rendered video: serves video
review/app.py - CLI pipeline → YouTube uploader: uploads if approved
pipeline.py - YouTube uploader → YouTube: publishes video and metadata
uploader/youtube.py - Web GUI → Topic research: controls generation
README.md - Web GUI → YouTube uploader: supports publishing
README.md