ffmpeg-skill
ffmpeg-skill is an Agent Skill for Claude Code, Cursor, Codex and any agent that reads SKILL.md, plus an MCP server for other clients. It teaches the agent a fixed workflow of probe, edit losslessly where possible, check and verify, and ships 42 tools that call ffmpeg and ffprobe: cut, join, silence and filler removal, fit to duration and aspect, captions with karaoke highlighting, overlays and motion graphics, HDR to SDR and LUTs, audio clean-up, sync with drift correction, multicam, loudness and delivery checks.
It is deliberately local and dependency-light: no cloud, no API keys and only the Python standard library. Transcription is optional and runs on your machine with Parakeet or Whisper when installed. Over MCP, tools/list advertises a core 12 tools by default to keep client context small, while all 42 stay callable by name, and a machine-readable contract describes every tool.
npx ffmpeg-skill installs the skill (with flags for Cursor, Codex or all three), and doctor checks the environment. The project publishes 66 before-and-after demos rebuilt from synthetic footage and tests against FFmpeg 5.1 through 9 in CI. License is MIT.
Features
Why choose it
Trade-offs
Maintenance check
- Last commit
- 2026-10-21
- Commits, last 90 days
- 100+
- Releases, last 90 days
- 30
- Latest release
- v2.10.8 (2026-10-21)
- Open issues
- 2
- Recent issues with at least one comment
- 3/11
Status rule: archived on GitHub = Archived; last commit within 90 days = Active; 91 to 365 days = Slowing; older = Stale. All figures come from the GitHub API on the date shown.
Most discussed open issues
- #143 Real-hands verification: Windows, Cursor and Codex installs on machines that are not the maintainer's, plus three editor reviews1 comments · 2026-09-11
Issue titles are quoted from GitHub. Open the link to check whether a problem still applies to the version you use.
Compatibility
Quick start
Use cases
Alternatives
Comments
No comments yet. Be the first!