This server connects Claude to Mixmake's transcript-based audio editing API. You can send audio files for transcription, then edit them by referencing specific word IDs in the transcript rather than timestamps or waveforms. When you're done cutting, rearranging, or removing words, you export the edited audio. Reach for this when you need to programmatically edit podcasts, interviews, or voice recordings through conversational edits like "remove all instances of 'um' and 'uh'" or "cut everything between word 45 and word 78." It's built for workflows where you want to edit audio the way you'd edit text, letting Claude handle the word identification and removal instructions while Mixmake processes the actual audio manipulation.
claude mcp add --transport http io.github.swappysh-mixmake https://mixmake.com/api/mcp