Skip to main content
Welcome to Luminy — a local-first desktop app that transforms any large language model into an active coding partner. Luminy gives your chosen AI direct access to your codebase, file system, shell, and git tooling, so it can read, write, search, and reason about your project just as a human collaborator would. Whether you prefer a private local model running entirely on your machine or a powerful cloud provider, Luminy keeps you in control.

Get Started

Install Luminy, connect a provider, and start chatting with your codebase in under 5 minutes.

Providers

Explore the 20+ supported AI providers, including local Ollama models and major cloud APIs.

Tools

Learn about the built-in agentic tools: file read/write, shell commands, git, and code search.

MCP Servers

Extend Luminy with Model Context Protocol servers to connect any external data source or API.

​
Getting up and running

You can go from a fresh install to a working AI coding session in three steps.
1

Download & Install

Visit luminy.tech to grab the installer for your operating system — macOS, Windows, or Linux. Run the installer and launch Luminy; no additional runtimes or dependencies are required.
2

Add a project

Click the + button in the sidebar and select the root folder of any codebase on your machine. Luminy indexes your project using Tree-sitter so the AI can navigate and understand your code immediately.
3

Start chatting with your codebase

Type a question or task in the chat input — for example, “Explain the architecture of this project” — and watch Luminy’s agentic loop explore your files, run searches, and produce a grounded, accurate answer.
Luminy is local-first: all your sessions, project data, and settings are stored locally on your own machine. Nothing is sent to Luminy’s servers. The only outbound traffic is the requests you authorize to your chosen AI provider.