C — Moderate risk signals Graded October 17, 2026Several concerns worth a closer look before you install.Based on public adoption and transparency signals as of Oct 17, 2026. Our opinion, not a code audit or security guarantee.
AI agent skill · Developer tools, Files & documents
minutes
Open-source, local-first Granola/Otter alternative that Claude Code, Codex, Cursor, and any MCP client can query. Meetings, calls, and voice memos transcribed on-device into markdown you own.
Bottom line: The tool itself isolates data locally, but you still expose transcripts to any AI agent you connect, and dependencies or integrations could be attack points.
- Popularity score
- 36/100
- Rank
- #434 of 500
- GitHub stars
- 1.6k
What it does
Minutes transcribes meetings, calls, and voice memos directly on your computer into markdown files you control. It works as an open-source alternative to services like Granola or Otter, and can connect to AI agents through MCP (Model Context Protocol) so they can search and reference your transcripts.
Who it’s for
Teams or individuals who want to keep meeting transcripts private and locally stored rather than uploading them to cloud services. Good for developers using Claude, Cursor, or other AI tools that need to search meeting notes.
Copy-paste setup
Check these steps against the project’s own instructions before running anything.
Follow the project's README for installation and MCP integration steps specific to your platform and AI client.
Risks
- Local storage access: The tool processes and stores all audio files and transcripts on your device; compromised access could expose sensitive meeting content.
- Transcript data leakage: If you later share transcripts with an AI agent, that agent or its provider could retain copies of sensitive discussions.
- Audio processing: It runs transcription locally but requires dependencies; malicious code in those dependencies could intercept audio before processing.
- MCP client trust: Any AI agent you connect via MCP can query your transcript database; a compromised agent could extract all meeting notes.
How to stay secure
Least-privilege permissions
- Run transcription with minimal OS permissions; avoid admin-level access.
- Connect only trusted AI clients via MCP; review what they request.
Network isolation
- Keep the tool offline or behind a firewall if transcripts contain highly sensitive data.
- Do not auto-sync transcripts to cloud storage.
Monitoring
- Audit which AI agents or users access your transcript database.
- Review dependency updates before upgrading.
Red flags to watch for
- Unexpected network activity or requests to phone home.
- Requests for permissions beyond audio processing and file storage.
- Transcripts appearing in unexpected locations or being uploaded without your action.
Alternatives
- Granola: Cloud-based meeting transcription with AI integration (less privacy).
- Otter.ai: Hosted transcription service with searchable archives (cloud-dependent).
- Whisper (OpenAI): Open-source speech-to-text model; requires your own setup.
- Notta: Privacy-focused local and cloud transcription option.