Highlight video
One YouTube-ready video of the length you choose, with chapters.
Hours of stream in. The best minutes out.
VOD Clipper turns long streams and podcasts into highlight videos, clips and ready-to-post Shorts. Claude picks the moments; your computer cuts them from the original, in full quality.
Download for Windows
Open source · View the code on GitHub
Windows only for now. Open this page on your PC to download.
See how it works ↓
- Runs on your computer
- Your videos never get uploaded
- Free
- Spanish & English
The problem and the idea
The problem
A 3-hour VOD is 6 GB. Claude.ai takes up to 500 MB per file and can't watch video.
The idea
VOD Clipper sends Claude a 15 MB pack it can read (transcript, thumbnails, energy signals) and cuts the clips from your original file.
6 GB→15 MB→full-quality clips
How it works
-
1 Pick a VOD and what you want
A highlight video, separate clips or Shorts. The app analyzes the recording on your computer.
-
2 Send the pack to Claude
Copy, attach, open: three clicks. Claude reads the transcript and thumbnails and picks the moments.
-
3 Review and get your files
Preview each moment, adjust it, then cut in seconds from the original.
Features
Separate clips
Each moment as its own clip, lossless.
Shorts auto-edit PRO · free in preview
Vertical 9:16, face framing (stacked for two people), word-by-word captions and a hook title.
Learns your style
Rate and trim, and the next jobs come out closer to how you edit.
Your videos stay on your computer
- Analysis and cutting run locally, with the app's own FFmpeg and Whisper. No GPU needed.
- Only two things touch the internet: the claude.ai chat you open yourself, and a one-time speech model download.
- No account with us, no tracking. This site has no analytics either.
Open source, so you can check it
The full code is public under the MIT licence. Anyone can read exactly what the app does with their videos.
- Read every line: no hidden uploads
- Builds made by a public GitHub workflow from that code
- Check your download: SHA-256 below
- Third-party licences listed (FFmpeg is GPL, source linked)
Your computer · everything on 127.0.0.1
Electron windowthe app you see
Local Python workerFastAPI, answers only the app
ffmpeg + Whispercuts and transcribes your files
claude.aiin your browser, opened by you, with the small pack
Requirements & install
- Windows 10 or 11, 64-bit
- About 2 GB of free disk space
- A free Claude account at claude.ai
- The first analysis downloads a ~464 MB speech model, once
Windows may say “Windows protected your PC”. The app isn't code-signed yet: click More info → Run anyway.
FAQ
Is it free?
Yes. While VOD Clipper is in preview, every feature is free, including the Pro ones like Shorts.
Do my videos get uploaded?
No. Your videos are analyzed and cut on your computer. Claude only gets the small pack you send it yourself: a transcript, thumbnails and loudness signals.
Do I need a paid Claude plan?
No. It works with a free Claude account at claude.ai. The pack is about 15 MB, well under the chat's upload limit.
Mac or Linux?
Windows only for now. The code is open, so a Mac or Linux build is possible later.
How long does it take?
On a typical laptop CPU, analyzing a 3-hour stream takes roughly 20–40 minutes, mostly transcription. Cutting the final clips takes seconds to a few minutes.
Which languages?
The app is in Spanish and English. Transcription works for most languages Whisper supports.
Is it open source?
Yes, under the MIT licence. It bundles FFmpeg (GPL v3), whose source is published next to every release. The code is on GitHub.
The best minutes of your stream, today.
Download for WindowsBehind the build
How to build a project like this →- Local-first desktop app: Electron + Python (FastAPI)
- ffmpeg + Whisper on the CPU, no GPU needed
- Evidence-based clip selection (loudness, laughter, speech rate)
- 200+ automated tests