Repolode
Submit
browser-use

browser-use/video-use

A Claude Code skill that edits raw video footage into a final cut from a chat prompt, using audio transcripts instead of frame analysis.

Open-source alternative to:Descript

View on GitHubWebsite
Stars
30,691
Forks
3,560
Open issues
164
Last commit
last month

What is it?

video-use is an open-source skill for coding agents such as Claude Code. You place raw footage in a folder, describe the edit in chat, and the agent produces a final.mp4. It works across content types like talking heads, tutorials, travel and interviews without requiring presets.

The agent reads the video through a word-level audio transcript from ElevenLabs Scribe, which includes speaker labels and audio events, rather than watching frames. It uses this to remove filler words and dead space, apply color grading, add short audio fades at cuts, and burn in styled subtitles. Animation overlays can be generated with HyperFrames, Remotion, Manim or PIL, and the rendered output is checked at each cut boundary before being shown.

Setup requires ffmpeg and an ElevenLabs API key, and the repo includes an install guide and a manual install path. Outputs are written to an edit/ subfolder next to the sources.

Key features

  • 01Removes filler words, false starts and dead air between takes
  • 02Auto color grades segments with presets or custom ffmpeg filters
  • 03Adds 30ms audio fades at every cut to avoid pops
  • 04Burns customizable subtitles, defaulting to two-word uppercase chunks
  • 05Generates animation overlays via HyperFrames, Remotion, Manim or PIL

Use cases

  • Turning a folder of raw talking-head takes into a polished short video
  • Assembling tutorial or interview footage into a cut without manual timeline work
  • Adding burned-in subtitles and animated titles to social media clips

Quick start

  1. 1

    Clone the repository

    git clone https://github.com/browser-use/video-use ~/Developer/video-use
  2. 2

    Install Python dependencies

    cd ~/Developer/video-use && uv sync
  3. 3

    Install ffmpeg (required)

    brew install ffmpeg
  4. 4

    Create the .env file, then add your ElevenLabs API key to it

    cp .env.example .env

Things to know

  • Requires an ElevenLabs API key for transcription.
  • Requires ffmpeg and an agent with shell access, such as Claude Code or Codex.

In collections

Claude Code tools, skills & pluginsAI agent skill librariesAI video & audio tools

Topics

#video-editing#claude-code#agent-skill#ffmpeg#speech-to-text#subtitles#automation

FAQ

What is video-use?
A Claude Code skill that edits raw video footage into a final cut from a chat prompt, using audio transcripts instead of frame analysis. video-use is an open-source skill for coding agents such as Claude Code. You place raw footage in a folder, describe the edit in chat, and the agent produces a final.mp4. It works across content types like talking heads, tutorials, travel and interviews without requiring presets.
Is video-use free to use? What license does it have?
video-use is open source under the MIT license. Check the license terms before using it in commercial products.
Is video-use still maintained?
Updated within 6 months. The last commit to video-use was on Oct 8, 2026.
What can I use video-use for?
Turning a folder of raw talking-head takes into a polished short video Assembling tutorial or interview footage into a cut without manual timeline work Adding burned-in subtitles and animated titles to social media clips
How do I install video-use?
Clone the repository: git clone https://github.com/browser-use/video-use ~/Developer/video-use Install Python dependencies: cd ~/Developer/video-use && uv sync Install ffmpeg (required): brew install ffmpeg Create the .env file, then add your ElevenLabs API key to it: cp .env.example .env
Where can I get video-use?
The source code and installation instructions are on GitHub: https://github.com/browser-use/video-use

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

Related projects