中文

AI Runner: private AI companion and layered canvas

A private AI companion and a layered canvas for AI art generation, running on your machine by default.

Verified

Core facts

Read by a script from the official API, not rewritten by a model.

License
GPL-3.0
Languages
Python
Maintainer
Capsize-Games
Verified

What it is

AI Runner is a private AI companion you shape (name, personality, voice, memory) and a layered canvas for AI art generation. Everything runs on your machine by default. AI Runner is free and GPL-3.0. Neither experience requires an internet connection, an API key, or a subscription.

Main features

  1. AI Companion: shape a named, voiced companion with persistent personality, shifting mood, and long-term memory built from your conversations.

  2. Layered Canvas: draw, paint, generate, and filter on a multi-layer canvas, convert sketches to images, composite scenes, and iterate in place.

  3. Image Generation: offers Z-Image Turbo with embeddings, image-to-image, inpainting, post-process filters, and background removal.

  4. Voice Conversation: provides full TTS and STT, so you can speak to your companion and hear it respond in a voice you choose.

  5. The companion builds long-term memory of you across sessions with RAG-powered recall.

Getting started

  1. The desktop download requires no Python setup.

  2. The advanced Python installation requires Python 3.13+.

  3. AI Runner downloads essential TTS/STT models automatically.

  4. LLM providers are Local (GGUF / llama.cpp), Ollama, OpenRouter, and OpenAI.

  5. Local LLM inference runs GGUF through a llama.cpp sidecar.

From the README · bash
xhost +local:docker && docker compose run --rm airunner
From the README · bash
   # torch family pinned to the CUDA 12.9 wheel line (issues #2036/#2041)
   pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu129
   # Desktop GUI; pulls in airunner-services==6.1.3 automatically
   pip install "airunner"
   # Model runtimes (LLM, STT, art, TTS); `headless` aggregates the four extras
   pip install "airunner-services[headless]"
   # Optional: native launcher and bundle tooling (also pulls the GUI)
   pip install "airunner-native[gui]"

Who it is for

It is private by default with no telemetry, and no data leaves your machine unless you explicitly enable an optional cloud feature. Model management includes built-in HuggingFace and Civitai downloaders with support for multiple local LLMs and image models. Windows support is experimental and community-maintained; Linux is the primary supported platform.

AI Runner can run as a headless HTTP API server, enabling remote access to LLM, image generation, TTS, and STT capabilities. AI Runner's local server uses HTTPS by default. The Ollama mode option runs as an Ollama replacement on port 11434.

The container publishes the headless daemon's HTTP API on port 8080. The OpenAI-compatible chat completions endpoint provides chat completion with tool support. AI Runner includes a best-effort content-safety filter for prohibited content.

The filter is defense-in-depth, not a guarantee. With no policy data loaded, the input-side filter is inactive and prompts pass through. The src package is the desktop GUI client, daemon client bridge, widgets, and application entry points.

The services package is the headless daemon, FastAPI server, runtime registry, orchestration, downloads, and persistence.

Sources

This page is written from the official sources below. Open them to check the original text.

Verified ·

This page is written by AI from the official sources above and checked automatically. This site has not tested it. Found a mistake? Tell us.