Inspiration
What it is
A set of command line tools for the unglamorous part of making a music video: knowing the song before you edit to it.
- BPM and key detection for any audio file
- Beat-synced shot lists: turn a track into a list of cuts that land on the beat
- Spotify Canvas maker: cut a 3 to 8 second vertical loop from an image or video
It also ships an MCP server, so an AI agent can call the same tools instead of guessing at timings.
Inspiration
Every music video starts with the same three questions: how fast is the song, where does the chorus start, and which moment should the video use. Answering them by ear takes an afternoon; the tools answer them in seconds.
How it works
The CLI decodes the audio locally, detects tempo and key, then writes a shot list as JSON so it can feed any editor or pipeline. Nothing is uploaded and no account is needed. The MCP server exposes the same three functions to agents.
Challenges
Tempo detection is easy on a four-on-the-floor track and hard on anything with a swing feel or a slow intro. We ended up combining an onset-based estimate with a manual override, so the number is always correctable.
What's next
The same team builds AIMusicVideoGen, a web app that turns a song and one photo into a lip-synced music video with every cut on the beat: https://aimusicvideogen.com
What it does
How I built it
Challenges I ran into
Accomplishments that I'm proud of
What I learned
What's next for music-video-cli
Built With
- cli
- mcp
- node.js
- spotify
- typescript
