Building VibeCast: Auto-Editing Podcast Videos with Python + FFmpeg
What if podcast video editing could be automated? 🎬
In this video, I share the concept for VibeCast, a proof-of-concept tool that automatically edits multi-camera podcast recordings. Using 3 cameras (two close-ups + one wide shot), screen capture, Python, and FFmpeg, the goal is to:
→ Record multiple video + audio tracks
→ Auto-sync them using a simple clap detection 🎤
→ Detect who’s speaking (via Whisper transcription)
→ Dynamically cut between speaker view, reaction shots, and wide angle
→ Export a final, ready-to-publish video in one command
This is not just about editing podcasts — it’s about vibe coding: proving that complex creative tools can be built fast, with minimal setup.
🚀 What you’ll see in this session:
→ The multi-camera podcast setup with DJI Osmo Pocket 3
→ Assumptions and challenges (mic syncing, clapboard detection, transcription)
→ The workflow from raw MP4 files → Python script → Final Cut MP4
→ How FFmpeg powers the entire pipeline
This is the first step in building VibeCast — an AI-assisted editing tool I’ll actually use for my own podcasts. Future versions may include subtitles, random wide-angle switching, and smarter speaker detection.
👉 Subscribe if you’re curious about vibe coding, AI tools for creators, and building practical automation for video workflows.
#vibecoding #vibecast #ffmpeg