HomeAboutProjectsRadarUniverseSkillsAchievementsContact
🌙
Back to the radar

Long-Video-to-Clips Open Source Stack

instagramstrong
Original post

TL;DR

A five-tool open-source pipeline combining WhisperX, PySceneDetect, Auto-Editor, Remotion, and FFmpeg to automatically convert long-form video into short clips. Automates the manual process of clipping long-form content such as podcasts and interviews into short-form video clips suitable for social media.

What it actually is

  • What: A five-tool open-source pipeline combining WhisperX, PySceneDetect, Auto-Editor, Remotion, and FFmpeg to automatically convert long-form video into short clips.
  • Who built it / maintained by: Multiple independent authors: Max Bain (WhisperX), Brandon Castellano (PySceneDetect), WyattBlue (Auto-Editor), Remotion team (Remotion), FFmpeg community. Presented by Marc Kaz.
  • Status: stable
  • Why it matters: Automates the manual process of clipping long-form content such as podcasts and interviews into short-form video clips suitable for social media.
  • How it compares to alternatives:
  • Opus Clip
  • Descript
  • Munch
  • Riverside.fm clips
  • Adobe Premiere Pro
  • GitHub stars: 25,269 · License: BSD-2-Clause · Archived: no

Links

  • Repo: https://github.com/m-bain/whisperX

Research evidence

General descriptions are reported, not independently tested. source_verified means the exact quotation was found in retrieved source text; it does not establish that an upstream claim works locally.

  • Identity: resolved · Basis: source_reported
  • reported: WhisperX is used for transcription in this stack (evidence: none)
  • reported: PySceneDetect is used for scene cut detection (evidence: none)
  • reported: Auto-Editor is used for silence removal (evidence: none)
  • reported: Remotion is used for programmatic video rendering (evidence: none)
  • reported: FFmpeg is used for final encoding (evidence: none)
  • reported: The stack can turn podcasts, interviews, and long videos into short clips automatically (evidence: none)
  • github-metadata · metadata · retrieved 2026-11-14T19:21:22.122Z · published unknown

Source excerpt: {"github_stars":25269,"open_issues":218,"last_pushed":"2026-11-12T17:34:05.000Z","license":"BSD-2-Clause","archived":false}

Kickstarter guide

Pseudocode / untested proposal. Check current upstream instructions and the target project before execution.

This is pseudocode only — verify all install commands and APIs from each repo before use. Step 1: run WhisperX on the source video to obtain a word-level timestamped transcript. Step 2: run PySceneDetect to identify scene boundaries. Step 3: pass the video through Auto-Editor to remove silent segments. Step 4: score transcript segments for clip-worthiness, then render selected ranges with captions using Remotion. Step 5: pipe the output through FFmpeg for final encoding to MP4.