anarlog is a privacy-first AI meeting notepad that captures and transcribes conversations locally on your device without joining the call as a bot.
The tool solves the problem of taking meeting notes with AI while keeping sensitive conversations private and off cloud servers. It captures audio directly from your device rather than joining as a participant, runs transcription locally on supported systems using built-in models, and stores all data in a local SQLite database with plain file attachments. Users can bring their own AI provider through API keys or run local models like Ollama, or opt into the tool's hosted AI services. The approach is designed for organizations that need to pass security reviews and individuals who want AI assistance without surrendering call recordings to third parties.
Adoption suits teams and individuals with strict data privacy requirements, companies with security review processes, and anyone uncomfortable storing meeting audio in the cloud. The tool explicitly positions itself as an alternative to Granola, offering the same meeting note functionality with local-first architecture as the distinguishing factor. Enterprise users have access to source-visible commercial components alongside the open-source community application. Cloud features like sync and sharing are entirely optional and off by default, so the tool functions completely offline if preferred.
The project maintains the open-source community application as a stable offering while the team focuses development efforts on a separate commercial product. The codebase is built with TypeScript, React, Rust, and Tauri, combining web technologies with native desktop capabilities. The tool provides export to Markdown to support integration with existing workflows.
