AI Runner is an open-source offline inference engine and desktop app for AI art, voice conversations and LLM-powered chatbots. Hosted on GitHub with about 1,300 stars under the GPL-3.0 license and actively developed through 2026, it runs everything on your own machine by default, with privacy as a core design goal.
AI Runner combines several local AI workloads in one application. It is a private AI companion you shape yourself — name, personality, voice and memory — and at the same time a layered canvas for AI art generation powered by Stable Diffusion. Real-time voice conversations are supported through local speech-to-text and text-to-speech, and LLM chatbots and automated workflows run against local models rather than cloud APIs. Because inference happens on-device, your prompts, images and conversations stay on your hardware. You can install it as a desktop download with no Python setup, run it from Docker images (including headless variants), or install it from PyPI if you prefer working from source. Desktop installers are published through itch.io, and the container images include Linux desktop and headless variants, so the same engine can also drive automated workflows on a server without a display.
AI Runner is free and open source under the GPL-3.0 license. Downloads and the source code cost nothing, and since models run locally, there are no API bills or subscriptions — you only need a capable computer or GPU.
Creators and hobbyists who want AI art, voice chat and a personal AI companion without sending their data to the cloud, and tinkerers who enjoy running open models locally with full control.