Local LLM
starnet
Pixel-art desktop where your local-first AI agents visibly do real work.
★ 1.0k starsMIT
View guide →Offline inference engine for art.
airunner is an open-source local llm project by Capsize-Games (Capsize-Games/airunner on GitHub). It has 1,317 GitHub stars and 104 forks, is released under the GPL-3.0 license, and was last updated on GitHub on 4 October 2026.
Decide what you want airunner to do first — offline inference engine for art, live voice conversations, LLM powere.
Pull the latest changes to stay current with fixes.
The README documents setup; start with the install section.
Change the config to match your workflow, then run it daily.
Working ai art support built in.
Working art support built in.
Working asset generator support built in.
Working chatbot support built in.
Readable code that is easy to modify.
Free to use; derivatives stay open.
Download the source and move into the project folder.
git clone https://github.com/Capsize-Games/airunner.git
cd airunnerInstall the Python dependencies from the project.
pip install -r requirements.txtLaunch with the project entry point — check the README for the exact command.
python main.py # see README for the exact entry pointYes, but note the copyleft: it is GPL-3.0 licensed, so distributed derivatives must stay open.
No. It is a self-run project — clone it and run it on your own machine. Some features may optionally call external APIs; check the README.
Yes, that is the point. The code is public, so you can run, modify, and host it yourself.
It is released under GPL-3.0. Read the LICENSE file in the repo for the exact terms.
Yes — it lives on GitHub at Capsize-Games/airunner. Open an issue or pull request; the contribution guide in the repo explains the process.
Pixel-art desktop where your local-first AI agents visibly do real work.
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
Local code-intelligence MCP server and CLI for coding agents.
TokenSpeed — a speed-of-light LLM inference engine designed for agentic workloads, with TensorRT-LLM-level performance and vLLM-level usability
Lightweight inference engine for running LLMs efficiently on any device.
Alibaba's lightweight, blazing-fast inference engine for on-device LLMs and edge AI.
The most-starred other Local LLM projects in the LibreAI catalogue — the realistic shortlist if airunner is not the right fit.
The simplest way to run open language models on your own machine.
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
LLM inference in C/C++
A high-throughput and memory-efficient inference and serving engine for LLMs
🤯 LobeHub is your Chief Agent Operator, organizing your agents into 7×24 operations by hiring, scheduling, and reporting on your entire AI team.
Free and open — grab the source, star the repo, and start creating.