This server gives Claude direct access to your macOS Voice Memos database, reading from the CloudRecordings.db file in your Group Containers folder. You get five tools: list_voice_memos for browsing recordings with search and pagination, get_voice_memo for metadata, get_audio to retrieve file paths or base64-encoded audio, get_transcript to extract Apple's transcripts from the custom tsrp atom embedded in m4a files, and transcribe_memo to generate new transcripts using SFSpeechRecognizer. Requires Full Disk Access permission and macOS Sonoma or later. Most useful when you want Claude to search through old recordings, pull transcripts into context, or coordinate with other audio processing tools like whisper-mcp for local transcription workflows.
claude mcp add --transport stdio jwulff-apple-voice-memo-mcp uvx apple-voice-memo-mcp