Radar check
What we found when we opened it. Every bar is one fact, with its date.
Last code pushed 28 Oct 2026, the same day before this check.
Latest release v0.12.0 on 23 Oct 2026.
Public since December 2024, a little over 1 year.
50,825 stars and 5,124 forks.
758 commits and 114 new issues in the last 90 days.
213 contributors. 654 pull requests merged in 90 days.
A security policy says how to report a problem. 3 published security advisories.
How the bars are scored
- Alive. Days since the last code push: a week or less 100, a month 90, three months 70, six months 50, a year 30, eighteen months 15, older 0.
- Releases. Age of the latest GitHub release, on a similar scale; no bar when the project publishes no releases.
- Maturity. Age of the project: eight years 100, five 90, three 70, two 55, one 40, six months 25, younger 10.
- Adoption. Stars on a log scale: 100,000 stars is 100, 10,000 is 80, 1,000 is 60, 100 is 40.
- Activity. Commits in the last 90 days: 300 or more 100, 100 85, 30 65, 10 45, any 25.
- Community. People who have contributed: 100 or more 100, 30 80, 10 60, 5 40, 2 25, one 10.
- Security. A published security policy 80, none 40; published advisories are listed, not scored.
Checked by Repo Radar on 28 Oct 2026 from the repository on GitHub.
In simple English
This program runs an artificial intelligence companion on your computer or phone. The companion has an animated face and talks with you using voice chat. You can connect it to games like Minecraft or chat apps like Discord. It works with many language model providers so you control the brain. The project is free to download and use on your own hardware.
What it is
AIRI is a TypeScript application that combines 3D character avatars, speech recognition, speech synthesis, and large language models. It runs as a desktop app, a mobile app, or inside a web browser.
What it is good for
Anyone who wants a personal virtual companion that can speak out loud, listen to their voice, and interact with games or chat rooms without sending private data to a closed service.
Summary (Key Points)from the README
Key features
- Multi-provider LLM support: Works with a long list of providers, including OpenAI, Anthropic Claude, Google Gemini, DeepSeek, Ollama, vLLM and OpenRouter, with AIHubMix marked as recommended.
- Multi-provider voice synthesis: Supports ElevenLabs, Microsoft/Azure Speech, OpenAI-compatible TTS, Alibaba Cloud Model Studio and local Kokoro TTS.
- Ears and speech input: Takes audio from the browser or Discord, with client-side speech recognition and talking detection.
- Avatar support: Controls both VRM and Live2D models, with auto blink, auto look-at and idle eye movement.
- Game playing: Listed as able to play Minecraft and Kerbal Space Program, and to play Factorio (work in progress, with a PoC and demo).
- Chat integrations: Lets the character chat in Telegram and Discord.
- In-browser storage: Offers in-browser database support through DuckDB WASM or pglite for memory.
- Native acceleration on desktop: The desktop version uses NVIDIA CUDA and Apple Metal by default, via the candle project.
- Web-based runtime: Runs in modern browsers and on mobile through PWA support, using WebGPU, WebAudio, Web Workers, WebAssembly and WebSocket.
Ways to install
| Platform or method | What it needs | Effort |
|---|---|---|
| Windows installer (x64 setup .exe) from v0.12.0 release | Windows x64 | Low: download and run |
| winget (`MoeruAI.AIRI`) | Windows with winget | Low |
| Scoop (airi bucket) | Windows with Scoop | Low |
| macOS DMG (darwin arm64) | Apple silicon Mac | Low |
| Homebrew Cask (`airi`) | macOS with Homebrew | Low |
| Linux from latest release page | Linux; package format not stated in README | Not stated |
| Browser version at airi.moeru.ai | A modern browser | None |
| Nix (`nix run github:moeru-ai/airi`) | Nix with flakes enabled | Medium |
| Run from source, browser (`pnpm dev`) | Repository checkout and pnpm | Medium |
| Run from source, desktop (`pnpm dev:tamagotchi`) | Repository setup per CONTRIBUTING.md, plus Electron libraries on NixOS | High |
| Run from source, mobile iOS (`pnpm dev:pocket:ios`) | Repository setup, an iOS device or simulator | High |
Important
- No official token: The project has no officially minted cryptocurrency or token, so the README asks readers to proceed with caution.
- Early stage: The README says development is at an early stage and the project is still seeking developers.
- Experimental mobile version: Stage Pocket is labelled experimental in the architecture diagram.
- Unfinished features: Co-playing Helldivers 2, Memory Alaya and pure in-browser local inference are still marked as incomplete.
- Missing providers: AWS Claude, Sparks and Volcano Engine are not yet supported and are marked PR welcome.
- Wireless mobile setup: Connecting the server channel to pocket in wireless mode requires running tamagotchi as root with sudo, and enabling secure websocket in settings.
- NixOS setup: Electron needs the FHS shell from `flake.nix` to find shared libraries.
- Browser limits: The README notes that features needing TCP or other non-Web technology, such as Discord voice or Minecraft, rely on the desktop side rather than the browser.
Written by an AI model from the project’s README; not part of the review, and it can be wrong.
- Supports both 3D VRM and 2D Live2D animated avatars that blink and look around.
- Works with dozens of different language model providers and local text-to-speech tools.
- Runs across web browsers, desktop operating systems, and mobile devices.
- Requires a powerful computer or paid external API keys to run smooth voice and vision models.
- Complex setup if you want to connect it to external games and chat bots.
- Early-stage software with frequent updates and evolving features.
Closed-source AI companion platforms and character chat apps
Ollama
- semperai/amica1.6k stars
An alternative browser-based interface for chatting with AI avatars using VRM models.
- elizaOS/eliza20k stars
A framework for building autonomous AI agents that connect to chat platforms and games.
- t41372/Open-LLM-VTuber-Lab47 stars
An open-source virtual YouTuber project powered by language models and voice synthesis.
- vllm-project/vllm96k stars
A serving engine that makes one graphics card answer many requests at once.
- infiniflow/ragflow93k stars
A question-answering system built on a document collection, with the sources shown beside every answer.
- basicmachines-co/basic-memory4.3k stars
Gives an AI assistant a memory that is stored as plain text files you can read.
brew install --cask airi