
Note67 - Private Meeting Notes Assistant
Open source, local-first AI meeting assistant

About Note67 - Private Meeting Notes Assistant
Introduction to Note67 - Private Meeting Notes Assistant
Note67 is an open-source, local-first meeting notes assistant for macOS. It captures audio from your microphone and system output, transcribes speech on-device using Whisper, and generates meeting summaries and action items with local large language models via Ollama. All processing occurs locally; no audio or text is sent to external servers.
The app is designed for users who need accurate, privacy-preserving meeting documentation, including remote teams, consultants, researchers, and security-conscious organizations. A clean interface supports real-time transcription, with options to pause and resume capture and review notes in light or dark mode.
Key Takeaways
- On-device transcription powered by Whisper; no cloud processing
- Local LLM summaries and action items generated via Ollama
- Real-time transcription with live text display
- Speaker distinction using combined microphone and system audio
- Echo handling with VAD and deduplication to reduce feedback artifacts
- Zero cloud dependencies: no accounts, subscriptions, or data collection
- Configurable models and preferences in a dedicated settings panel
- Native macOS app built with Rust and Tauri for performance
How Note67 - Private Meeting Notes Assistant Works
Note67 captures both microphone input and system audio to record meetings and calls. Using whisper-rs, it transcribes speech locally so that audio and transcripts remain on the device. Transcripts appear live during a session, and users can pause and resume recording as needed.
When enabled, the app sends the local transcript to Ollama, which runs a local LLM to produce structured outputs such as meeting summaries, minutes, and action items. This workflow maintains privacy by keeping all processing on the device and storing data locally. Echo handling and voice activity detection help reduce duplicate or echoed text when speakers are not using headphones.
| Component | Technology |
|---|---|
| Audio capture | ScreenCaptureKit; mic + system audio |
| Transcription | whisper-rs (Whisper models), on-device |
| Summarization | Ollama with local LLMs |
| Interface | React, TypeScript, Tailwind CSS |
| Runtime | Rust, Tauri v2 |
| Storage | SQLite |
| Noise/echo control | VAD and deduplication |
| Platform | macOS 13+ |
Core Benefits and Applications
- Privacy-preserving documentation: All audio and text stay on the device, suitable for confidential meetings, legal reviews, and internal planning.
- Reliable note-taking for remote work: Real-time transcripts support video calls and webinars without relying on external services.
- Structured follow-ups: Generate summaries, minutes, and action items to streamline team alignment and task tracking.
- Research and interviews: Capture interviews, user research sessions, and stakeholder calls with speaker distinction and clear transcripts.
- Consistent records: Local storage and configurable models promote repeatable, auditable note-taking workflows without external dependencies.