LumaBrowser
Local AI desktop app: a Chromium browser with a GPU-aware llama.cpp runtime, OpenAI-compatible API, AI chat and agents, Stable Diffusion image/video, TTS and mu
What is LumaBrowser?
Local AI desktop app: a Chromium browser with a GPU-aware llama.cpp runtime, OpenAI-compatible API, AI chat and agents, Stable Diffusion image/video, TTS and music, plus an MCP server so Claude Code, Codex and Cline can drive your browser. Windows, macOS, Linux.
Overview of LumaBrowser
What it is
LumaBrowser is a local AI desktop application that combines a Chromium browser with a GPU-aware llama.cpp runtime, an OpenAI-compatible API, AI chat and agents, Stable Diffusion image and video generation, text-to-speech and music generation, and an MCP server. It is designed to run entirely on the user's own hardware, supporting Windows, macOS, and Linux. The application is free and open source under the AGPL-3.0 license.
Key features
- Integrates a Chromium browser with an ad blocker, Chrome extensions, a password manager, private windows, and a Shift Shift search for settings or chats.
- Provides local model runtimes (llama.cpp and MLX) for NVIDIA, AMD, Intel, Qualcomm, and Apple hardware, with the ability to split a model across GPUs or machines and use a drafter for speed.
- Supports image, video, music, and voice generation using models such as FLUX, Qwen-Image, MiniMax-H3, and Wan 2.2, along with a CivitAI browser and area editing/outpainting.
- Includes named agents, a coding agent that runs in the terminal and editor, tools the agent writes for itself, and the ability to turn signed-in sites into APIs.
- Offers computer use capabilities where the agent operates other desktop apps, shows which window it is in, and asks before accessing password fields or checkouts.
- Provides automation and sharing features including triggers, page watches, recurring tasks, and the ability to share models and GPUs with other devices or a live tab by link.
Use cases
LumaBrowser is suitable for running large language models, image generation, video generation, and voice models locally on personal hardware. It can be used for AI chat with web browsing, file access, a knowledge base, and custom agents. Developers can connect tools like Claude Code, Codex, Cline, or OpenCode to use a real browser via MCP, ACP, WebDriver, or CDP. It also enables desktop control that runs locally without per-step costs, with screenshots never leaving the machine.
Getting started
Download the appropriate file for your computer from the latest release on the repository page. For Windows, use LumaBrowser.Setup..exe or the portable version. For Windows on Arm, use LumaBrowser.Setup.-arm64.exe. For Mac (Apple Silicon), use LumaBrowser--arm64.dmg. For Linux, use LumaBrowser-.AppImage or -arm64.AppImage on Arm. Alternatively, with Node 18 or newer, run: sh npx lumabrowser start On first run, open the Models section to search for a model. LumaBrowser reads your hardware and shows a fit badge and predicted speed before download. Then open a chat and give it a task that requires the web.
License and pricing
The license is AGPL-3.0. The pricing type is open source. No pricing information is stated beyond the open source license.
Based on the project's public README and repository metadata.