Workflow — Podcast Pipeline

What: Take a raw multi-mic recording and ship a broadcast-clean, caption-complete, chapter-tagged, loudness-compliant episode to every podcast platform.

Skills used

media-whisper, media-demucs, media-denoise-ai, media-subtitle-sync, media-sox, ffmpeg-audio-filter, ffmpeg-audio-fx, ffmpeg-audio-spatial, media-ffmpeg-normalize, ffmpeg-detect, ffmpeg-metadata, ffmpeg-subtitles, ffmpeg-captions, media-batch, media-cloud-upload, media-imagemagick.

Pipeline

Step 1 — Capture + concat multi-source

Join separate mic files into aligned tracks via amix / amerge. Auto time-align against a clap or common audio marker.

Step 2 — AI denoise per stem BEFORE mixing