Generate meeting notes with a local LLM

Reki note can generate meeting notes by connecting to a local inference server running on your PC. With this approach, meeting audio and transcripts are never sent to an external AI service. This page walks through setup using the recommended app, LM Studio.

Last updated: September 29, 2026

What is a local LLM?

A local LLM is a large language model that runs entirely on your own PC instead of a cloud AI service.

Normal note generation sends text to a cloud AI, but with a local LLM the transcript and meeting notes stay on your machine. This suits sensitive meetings or cases where you want to avoid sending data externally.

Two ways to connect

In Reki note settings you can use a local LLM in two ways:

• Local LLM: Reki note downloads the model and runs it on your device. No extra app to install.

• Local endpoint: connect to an OpenAI-compatible server you installed separately, such as LM Studio or Ollama. Best if you want to choose the model yourself.

This page covers the flexible, recommended “local endpoint” approach.

Recommended app: LM Studio

LM Studio is a free GUI app for running local LLMs. It lets you search and download models and start an OpenAI-compatible server, all inside the app, which makes it a good starting point for beginners.

It supports both macOS and Windows, including Apple Silicon.

Download LM Studio

Setup steps

Follow these steps to use a local LLM from Reki note.

Default LM Studio server URL:

http://127.0.0.1:1234

1. Download and install LM Studio.
2. Open LM Studio, find a model in the search tab, and download one.
3. Open the Developer tab on the left and start the server (“Start Server”). The default port is 1234.
4. In Reki note, open Settings > Data processing and select the “Local endpoint” row.
5. Enter the server URL http://127.0.0.1:1234 (or use “Detect local server” to fill it in).
6. Press Refresh to list models, then choose the model to use.
7. Press “Test and confirm”. Once it succeeds, the local LLM is used for note generation.

LM Studio “Developer” tab > “Local Server”. When Status shows “Running” and Reachable at shows the server URL (http://127.0.0.1:1234 by default), the server is up.

Latest LM Studio REST API docs

Using Ollama or another app

If you already use Ollama or llama-server, you can connect by entering that server URL directly. The default Ollama port is 11434.

In any case, any server that exposes an OpenAI-compatible /v1/chat/completions endpoint will work.

Download Ollama

Troubleshooting

• No models listed: make sure the local server is running and the URL/port is correct.

• Connection test fails: check that a model is downloaded and the server exposes an OpenAI-compatible API.

• Slow responses: switch to a smaller model in LM Studio, or set “Reasoning (thinking)” to Off for faster replies.

All of these options are in Settings > Data processing.