At a Glance:
anarlog is an open-source, local-first AI meeting notetaker that records audio, transcribes on-device, and saves notes to markdown files, allowing users to bring their own LLM provider and own their meeting data.
Overview:
anarlog is a local-first meeting notetaker application designed to provide AI-powered meeting summaries without depending on a cloud backend. It records meeting audio, performs speech-to-text transcription directly on the user's machine, and uses a configurable LLM to generate notes. All output is saved as plain markdown files on the local disk, giving users direct control over their data. The application is MIT-licensed and can be used with various AI providers, from cloud APIs like OpenAI and Anthropic to local models running via Ollama or LM Studio, and can also be self-hosted by building it from the source.
Key Decision Points:
Data storage format: Meeting notes are stored as local markdown files, allowing users to inspect, search, and sync them with tools like Dropbox or git.
Audio processing model: Transcription is performed on-device, meaning raw audio does not leave the user's computer.
AI provider flexibility: Users must supply their own LLM credentials or configuration and can use any OpenAI-compatible service or local model.
Deployment and self-hosting: There is a pre-built release for direct download, and the project can also be cloned and built for self-hosting.
No account system: The application operates without user accounts, data collection, or a hosted backend.
Core Features:
Local recording and transcription: Records system audio and transcribes speech entirely on the local device.
Markdown note generation: Uses a configurable LLM to structure the transcript into meeting notes and saves them as markdown files on disk.
Bring Your Own LLM (BYOLLM): Supports connecting to OpenAI, Anthropic, Gemini, OpenRouter, Ollama, LM Studio, or any OpenAI-compatible API.
Local-first data control: Every meeting output is a standalone
.mdfile that users can manage, search, and synchronize with file-based tools.Self-hostable MIT license: The source code is open-source, forkable, and can be built and run independently.
Use Cases:
Developers managing meetings: A developer who wants AI-generated meeting notes stored as version-controllable markdown files alongside their project documentation.
Privacy-conscious professionals: Users in fields where sending meeting audio to third-party cloud servers is not acceptable and who require local transcription and data storage.
Local AI experimenters: Individuals running local LLMs through Ollama or LM Studio who want to use their own models to generate structured meeting summaries.
Open-Source Alternative Value:
anarlog provides an alternative to proprietary meeting notetakers by separating the tool's core function from a mandatory cloud backend. Its local-first architecture ensures data is saved as portable markdown files rather than locked in a proprietary cloud format. By combining on-device transcription with a bring-your-own-model approach, users can choose AI providers based on cost, privacy, or capability without giving the application itself access to their data or audio. Its MIT license and self-hosting option further allow for independent use and modification without reliance on a vendor's ongoing service.
