Who it is forA developer who prefers to see every install step instead of letting the agent do it.
The problemAgent-driven installs can be opaque about what changed on your machine.
The outcomeVideo Use registered as a skill in Claude Code or Codex, with dependencies and the API key in place.
What you need first
- Claude Code, Codex or another agent with shell access
- ffmpeg (required) and optionally yt-dlp
- Python with uv or pip
- An ElevenLabs API key for transcription
Steps
- Run
git clone https://github.com/browser-use/video-use ~/Developer/video-use - Symlink it:
ln -sfn ~/Developer/video-use ~/.claude/skills/video-usefor Claude Code - Install dependencies with uv sync (or pip install -e .) inside the folder
- Run
brew install ffmpegand optionallybrew install yt-dlp - Copy .env.example to .env and add ELEVENLABS_API_KEY
Prompt for your agent
Install browser-use/video-use manually for Claude Code following the README's manual steps, verify ffmpeg is on PATH, and tell me it is ready without transcribing anything.
Caveat.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for browser-use/video-use and "Install Video Use as a Claude Code skill by hand"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer