How to Use Anarlog: Private AI Meeting Notes with Your Own LLM

Anarlog promises private, local-first AI meeting notes — here is how to actually put it to work, from a quick desktop start to a fully self-hosted setup with your own LLM.

Step 1: Run the desktop app (easiest path)

Grab the latest release for your platform from the GitHub releases page. Open the app, join a meeting, and Anarlog records it while transcription runs locally on your device. Canonical meeting data lands in a local SQLite database — no account, no upload, no vendor lock-in.

Step 2: Connect your LLM of choice

Anarlog is model-agnostic. In settings, point it at OpenAI, Anthropic, Gemini, OpenRouter, Ollama, or LM Studio — or anything exposing an OpenAI-compatible endpoint. For a zero-cloud workflow, run Ollama locally and pick a local model; transcription and note generation then both happen entirely on your machine.

Step 3: Self-host from source

Clone the repo and build it, or use the included apps/api/Dockerfile for the API service. The repository also ships a compose setup under enterprise/control-plane if you need the control-plane components. System requirements are modest: 2+ CPUs and 4 GB RAM get you started comfortably, with 4 CPUs and 8 GB recommended for heavy transcription workloads.

Step 4: Export and share

When collaboration is needed, export Markdown or opt into the optional encrypted CloudSync and sharing features. You stay in control of when data leaves your device.