Repolode
Submit
Aseiel

Aseiel/VideoHighlighter

A free, offline desktop app that scores and explains highlight moments in local video using scene, motion, audio, object and transcript signals.

Open-source alternative to:DescriptTwelve LabsRunway

View on GitHub
Stars
255
Forks
33
Open issues
1
Last commit
yesterday

What is it?

VideoHighlighter is a local desktop app for finding the moments that matter in video you do not want to upload. You drop in raw footage, and it scores segments using scene changes, motion, audio, detected objects, actions and the transcript, then exports a highlight reel and individual clips. The analysis runs on your machine, and the basic pipeline needs no API key.

Each kept moment comes with a report that breaks down the per-signal points, shows what fired, and lists high-scoring segments that missed the cut. Claims drawn only from the transcript are marked as unmeasured rather than scored. Users can also write composition rules that combine detections, such as one object inside another or a count held over a time window, and name the resulting event.

The app runs on Windows, macOS and Linux, with support for Intel, AMD and NVIDIA GPUs. It includes an Auto tab that can ingest a camera card, cut highlights, build a reel and add music, plus Whisper-based transcripts and subtitles. It is licensed under AGPL-3.0.

Key features

  • 01Scores segments from scene, motion, audio, object, action and transcript signals.
  • 02Shows per-signal point breakdowns explaining why each moment was kept.
  • 03Lets you define composition rules that turn detection combinations into named events.
  • 04Exports highlight reels and individual clips.
  • 05Generates transcripts and subtitles with local Whisper.
  • 06Runs offline with support for Intel, AMD and NVIDIA GPUs.

Use cases

  • Condensing a long livestream or VOD into a short highlight reel.
  • Reviewing CCTV or trail-camera footage to find the minutes when something moved.
  • Turning a GoPro or drone card into an edited film with music.
  • Creating chapters and subtitles for long interviews, lectures or podcasts.

Quick start

  1. 1

    Install from source on Linux (FFmpeg included)

    pip install -r requirements.txt
  2. 2

    Launch the app

    python main.py

Things to know

  • Windows build is not code-signed, so you must click through a warning to run it.
  • macOS build needs a one-time quarantine flag removal before it opens.
  • Released under AGPL-3.0, so modified versions offered over a network must publish their source.
  • Some features like live object overlays and open-vocabulary detection are only in the separate paid Pro edition.

Topics

#video-analysis#local-ai#offline-ai#whisper#object-detection#video-search#subtitle-generation#ollama#action-recognition-dataset#cuda

FAQ

What is VideoHighlighter?
A free, offline desktop app that scores and explains highlight moments in local video using scene, motion, audio, object and transcript signals. VideoHighlighter is a local desktop app for finding the moments that matter in video you do not want to upload. You drop in raw footage, and it scores segments using scene changes, motion, audio, detected objects, actions and the transcript, then exports a highlight reel and individual clips. The analysis runs on your machine, and the basic pipeline needs no API key.
Is VideoHighlighter free to use? What license does it have?
VideoHighlighter is open source under the AGPL-3.0 license. Check the license terms before using it in commercial products.
Is VideoHighlighter still maintained?
Actively maintained. The last commit to VideoHighlighter was on Oct 18, 2026.
What can I use VideoHighlighter for?
Condensing a long livestream or VOD into a short highlight reel. Reviewing CCTV or trail-camera footage to find the minutes when something moved. Turning a GoPro or drone card into an edited film with music. Creating chapters and subtitles for long interviews, lectures or podcasts.
How do I install VideoHighlighter?
Install from source on Linux (FFmpeg included): pip install -r requirements.txt Launch the app: python main.py
Where can I get VideoHighlighter?
The source code and installation instructions are on GitHub: https://github.com/Aseiel/VideoHighlighter

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects