Record and transcribe meetings privately without any cloud subscription or account sign-up.
Save meeting notes as plain Markdown files you can search, version-control, or sync with any tool you already use.
Connect your own OpenAI, Anthropic, or local Ollama model to summarize or analyze meeting content privately.
Requires connecting your own AI provider (OpenAI, Anthropic, Gemini, or a local model via Ollama or LM Studio) before meeting notes can be generated.
anarlog is an open-source desktop application that takes notes during meetings using AI, without sending any audio or data to external servers. When you join a meeting, it records and transcribes your audio on your own device using local speech models. After the meeting, your notes are saved as a plain Markdown file on your hard drive, not in anyone's cloud. The application does not come with a built-in AI provider. Instead, you connect your own: OpenAI, Anthropic, Gemini, or any locally running model through tools like Ollama or LM Studio. This means you decide which AI processes your meeting content, and you can switch providers or use one that never touches the internet at all. Because every meeting is just a .md file on your disk, you can open, read, search, or back it up with any tool you already use. The README explicitly lists this as a design goal: you can cat it, grep it, sync it with Dropbox, iCloud, Syncthing, or git. There is no vendor lock-in because there is no central service to be locked into. No accounts are required and no tracking is built in. The project is MIT licensed, which means you can fork it, modify it, or use it as the basis for your own product. It is built with Tauri (a framework for making desktop apps using web technologies with a Rust backend), React, and TypeScript. It runs on standard desktop operating systems and can also be self-hosted by cloning the repository and building from source. The README notes that the same team is now focused primarily on a separate product called char. anarlog remains open-source and maintained, but active feature development has shifted elsewhere. The project was previously named Hyprnote before going through several name changes.
← fastrepl on gitmyhub — every repo by this author, as a profile.
Verify against the repo before relying on details.