Product Introduction
- Overview: Note67 is a privacy-first macOS application leveraging Whisper speech-to-text and Ollama local LLMs for on-device meeting transcription and summarization.
- Value: Enables professionals to generate searchable meeting records and AI insights without compromising sensitive data through cloud processing.
Main Features
- Local Transcription: Executes Whisper-powered speech recognition entirely on-device using whisper-rs Rust bindings, ensuring zero audio data transmission.
- Speaker Distinction: Identifies participant voices through dual-channel audio processing combining microphone input and system audio capture.
- Private AI Summaries: Generates contextual meeting summaries via locally-hosted Ollama LLMs with customizable open-source models.
- Echo Handling: Implements Voice Activity Detection (VAD) and deduplication algorithms to filter speaker echo during non-headphone use.
Problems Solved
- Challenge: Eliminates data privacy risks in confidential meetings where cloud-based transcription services violate compliance requirements.
- Audience: Security-conscious professionals in legal, healthcare, and enterprise sectors handling sensitive discussions.
- Scenario: Securely documenting board meetings where proprietary strategies or regulated information must remain on-premises.
Unique Advantages
- Vs Competitors: Unlike cloud-dependent tools, Note67 provides true zero-trust architecture with end-to-end local processing and no subscription requirements.
- Innovation: Combines Whisper's speech recognition with Ollama's local LLM orchestration in a Tauri-based Rust framework for native macOS performance.
Frequently Asked Questions (FAQ)
- Q: Does Note67 require internet connectivity? A: No, all processing occurs offline using local Whisper models and Ollama LLMs installed on your Mac.
- Q: How does speaker distinction work without cloud processing? A: By analyzing separate audio channels from microphone and system output using ScreenCaptureKit APIs for participant isolation.
- Q: Can I use custom Whisper or LLM models? A: Yes, Note67 supports model customization through Ollama's local LLM framework and whisper-rs configuration.
