
Note67 is a private meeting notes assistant that redefines how you capture and process meeting audio. It is an open-source, local-first AI tool designed for anyone who values privacy and control over their data. Unlike conventional cloud-dependent solutions, Note67 runs entirely on your device, ensuring no audio or transcript ever leaves your machine. The app combines real-time speech-to-text via Whisper with on-device AI summarization powered by Ollama, letting you generate intelligent meeting summaries without any internet connection. Built with Rust and React, it delivers native performance on Windows and macOS. For professionals, remote workers, and privacy advocates, Note67 offers a secure, efficient way to turn conversations into actionable notes while keeping sensitive information completely local. Its core promise is simple: your meetings, your data, your device.
Modern meetings are filled with critical information, yet taking manual notes is distracting and often incomplete. Cloud-based transcription services promise convenience, but they force you to upload audio to remote servers, creating potential privacy breaches and violating confidentiality policies. This is particularly problematic for legal, healthcare, and corporate settings where data governance is paramount. Note67 eliminates this trade-off by processing everything locally. There are no accounts to create, no subscriptions to manage, and no data collection of any kind. Once installed, the software works offline, meaning you never have to worry about third-party access to sensitive discussions. The pain point of balancing productivity with confidentiality is resolved entirely—Note67 ensures that you can capture every detail of a meeting without compromising on security. It is a privacy-first solution that finally respects the user's digital sovereignty.
At the heart of Note67 is its local transcription engine, powered by Whisper through whisper-rs. This feature converts spoken audio into text in real time without ever sending data off your device. The transcription process runs efficiently thanks to Rust’s performance and Tauri’s lightweight framework, ensuring minimal system resource usage. Users can watch transcripts appear live on their screen as they speak, with the ability to pause and resume at any time. This live transcription is invaluable for reviewing what has been discussed during a meeting or for instantly capturing action items. Because everything happens locally, there is zero latency from network calls, and the privacy guarantees are absolute. The integration with Whisper means you get state-of-the-art speech recognition accuracy while maintaining full control over your data. Note67’s local transcription makes meeting documentation seamless and secure.
admin
Note67 includes sophisticated audio processing to enhance clarity. Speaker Distinction automatically separates your voice from other meeting participants by leveraging microphone input and system audio capture. This dual-source approach allows the assistant to differentiate who is speaking, making meeting notes far more structured and readable. In practice, when you are in a virtual meeting, Note67 captures both your voice (through the mic) and the output of the meeting app (system audio), and it intelligently labels segments. Complementing this is Echo Handling, which uses smart Voice Activity Detection (VAD) and deduplication filters to eliminate speaker echo that can occur when using speakers instead of headphones. This ensures that repeated or feedback-created audio artifacts are not transcribed, preserving the cleanliness of the transcript. These audio-enhancement features make Note67 suitable for real-world meeting conditions, delivering accurate and noise-free records.
Once a meeting transcript is ready, Note67’s Private AI Summaries let you generate concise, intelligent summaries directly on your device using Ollama and any compatible local large language model. This feature eliminates the need for cloud-based AI services like OpenAI or Gemini, keeping the entire summarization workflow entirely private. You have full control over which LLM to use—be it a lightweight model for speed or a more powerful one for deeper analysis. The summaries highlight key points, decisions, and action items, turning lengthy transcripts into actionable meeting notes. With no data leaving your machine, even highly confidential discussions can be summarized safely. This local AI approach not only protects sensitive information but also avoids recurring API costs and internet dependency. Whether you are a developer experimenting with open models or a business user needing reliable meeting reports, Note67 delivers unmatched privacy without sacrificing intelligence.
Note67 operates as a native desktop application for Windows and macOS, built with Tauri v2 and React 19. After a one-time installation, it works entirely offline. The workflow is straightforward: you start a recording session, and Note67 accesses your microphone and system audio via ScreenCaptureKit (on macOS) or equivalent APIs. The audio feeds are processed in real time; whisper-rs performs speech-to-text, the VAD engine filters out silence and echo, and the output is stored locally in a SQLite database. Users can view the live transcription in a clean, minimal interface that offers light and dark modes. After the meeting, with a single click, you can generate an AI summary using your local Ollama setup. All configurations, from Whisper model selection to Ollama server preferences, are managed through a settings panel. The result is a seamless, self-contained system that respects your privacy at every step.
Consider a product manager in a confidential roadmap meeting. Using Note67, they can capture the full discussion with live transcription, automatically distinguish speakers, and later generate a private summary without ever exposing corporate secrets. A freelance consultant conducting client interviews can record sessions and instantly produce detailed notes for reference, all while assuring clients that no data is stored in the cloud. Developers who prefer open-source tools can integrate Note67 into their workflow, customize the LLM for domain-specific summaries, and even contribute to the codebase. Legal professionals who require offline documentation can use the app to securely archive meeting minutes. In each scenario, the outcome is the same: accurate, searchable meeting records with zero privacy trade-offs. Note67 empowers users to focus on the conversation, knowing that documentation is handled reliably and privately.
Note67 caters to privacy-conscious professionals, remote workers, developers, and anyone who handles sensitive meetings on Windows 10+ or macOS 13+. The application is completely free and open source, with no hidden costs, subscriptions, or account requirements. Its tech stack—Rust, Tauri, React, Tailwind CSS, SQLite, whisper-rs, and Ollama—ensures top-tier performance and modifiability. Because it is open source under a permissive license, users can audit the code, suggest improvements, or tailor it to their needs. The platform choice of Windows and macOS makes it accessible to most desktop users. In a world where data privacy is increasingly rare, Note67 stands out by offering a capable, local-first meeting notes assistant that doesn’t compromise on features. It proves that you don’t need the cloud to have intelligent meeting documentation; you just need the open source tools and the commitment to keep your data yours.
Note67 is designed for privacy-conscious professionals, remote and hybrid workers, open-source developers, legal professionals, healthcare practitioners, and corporate teams. It suits anyone who handles sensitive or confidential meeting content and requires on-device transcription and AI summarization. The tool is especially valuable for those on Windows 10+ or macOS 13+ who prefer self-hosted, offline-capable software and reject cloud-based subscription models.