Radar check
What we found when we opened it. Every bar is one fact, with its date.
Last code pushed 11 Oct 2026, 16 days before this check.
Latest release v2.79.0 on 4 Oct 2026.
Public since October 2023, about 3 years.
1,130 stars and 229 forks.
162 commits and 0 new issues in the last 90 days.
30 contributors. 39 pull requests merged in 90 days.
No security policy file says how to report a problem. No 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 tool helps you set up and run an animated character driven by artificial intelligence. You can chat with the character directly or have it respond to comments on live video streams. It connects to various language models and voice generators to bring the character to life. You need your own API keys for the services it uses. The project is free for personal use, but commercial use requires a separate licence.
What it is
A TypeScript web application that combines three-dimensional models, text-to-speech engines, and large language models into an interactive character interface. It includes features for live stream integration, screen capture commentary, and memory storage.
What it is good for
Solo creators and developers who want to launch an interactive AI character for live streams, digital signs, or personal desktop chat without building the interface from scratch.
Summary (Key Points)from the README
Key features
- Multiple AI backends: Supports 17 LLM options, including OpenAI, Anthropic, Google Gemini, Ollama, and LM Studio.
- Voice synthesis: Supports 11 speech engines, including VOICEVOX, AivisSpeech, Style-Bert-VITS2, and ElevenLabs.
- Character models: Loads VRM 3D models with motion-tag pose control, Live2D models (Cubism 3 or later), or motion PNGTuber characters.
- Memory: Keeps recent conversation turns and adds RAG-based long-term memory from past chats.
- Multimodal input: Recognizes camera video and uploaded images when generating replies.
- Live streaming: Reads YouTube live comments through the YouTube API or OneComme and replies automatically, with optional unprompted speech in conversation-continuation mode.
- Digital signage: Demo-terminal mode offers fullscreen display with passcode lock, NG-word filtering, and input length limits. Presence detection greets visitors via camera face detection.
- Advanced modes: Includes OpenAI Realtime API low-latency dialogue, audio mode, a reasoning display mode, and a game commentary mode that analyzes screen captures every so often.
- External integration: WebSocket links to server apps, a slide presentation mode, and a message-receiving API that lets outside systems make the character speak.
Ways to install
| Platform or method | What it needs | Effort |
|---|---|---|
| Local development (`npm install`, `npm run dev`) | Node.js 24.x, npm ^11.6.2, git | Moderate |
| Double-click launcher | Completed first-time setup; `LAUNCH.bat` on Windows or `LAUNCH.command` on macOS | Low |
| Docker Compose (`docker compose up -d`) | Docker and a `.env` file | Low |
| Vercel | Vercel account, GitHub import, API keys set as environment variables | Low to moderate |
| Cloudflare Workers | Cloudflare account, Wrangler CLI, OpenNext, and Wrangler secrets for server-side keys | Higher |
Important
- Commercial license: Since v2.0.0 the project uses a custom license. Commercial use requires a separate commercial license, while personal, educational, and non-profit use is free.
- Live2D restrictions: The bundled Live2D model cannot be used for image generation, AI input, reference, training, or conversion, and it cannot be modified or redistributed. Commercial use is also prohibited and is not covered by the AITuberKit commercial license.
- Server secret access mode: Public API use of server-side secrets such as `OPENAI_API_KEY` requires `AITUBERKIT_SERVER_SECRET_ACCESS_MODE`. Without it, some APIs return `403`.
- API key handling: Calls to AI and TTS services go through the backend, so API keys need careful management during deployment.
- Demo rate limit: `AITUBERKIT_DEMO_RATE_LIMIT_PER_MINUTE` is an in-process memory limit. Serverless or multi-replica setups apply it per instance, so production deployments should add a WAF or shared rate limiting.
- Cloudflare limits: Builds disable the filesystem API. Files over 25MB in `public/` and files with non-ASCII names are excluded from deployment. `NEXT_PUBLIC_*` values are embedded in client code at build time.
- Priority feature work: Paid priority implementation is quoted per feature and is separate from the commercial license. Implemented features are published as part of the open-source project.
Written by an AI model from the project’s README; not part of the review, and it can be wrong.
- Supports multiple character formats including 3D models and 2D Live2D files.
- Connects to a wide range of language models and voice generators.
- Includes ready-to-use modes for live stream comment integration and idle chat.
- Custom licence restricts commercial use without prior arrangement.
- Requires external API keys which incur ongoing usage costs.
- Initial setup requires a Node.js development environment.
Writing a custom web application from scratch to connect language models with animated character models and live stream chat feeds.
Node.js (a way to run JavaScript on your computer outside a web browser)
- spotify/pedalboard6.4k stars
Applies audio effects to recordings from Python, including studio plug-ins, without calling another program.
- iver56/audiomentations2.3k stars
Spoils sound recordings on purpose, 39 ways, so a model trained on them copes with real rooms.
- shaka-project/shaka-packager2.6k stars
Cuts a video file into the pieces and index files that a browser needs in order to stream it.
git clone https://github.com/tegnike/aituber-kit.git cd aituber-kit npm install ncp .env.example .env npm run dev