AITuberKit is an open-source building kit that lets anyone spin up a web app for talking with AI characters. Published in April 2023 by tegnike, an individual developer in Japan, it bundles VRM and Live2D character models, multiple speech synthesis engines including VOICEVOX, and automatic replies to YouTube Live comments. Clone the repository, set your API keys, and you have a working foundation for an “AITuber” — a character that speaks on screen and answers viewer comments. Note that the developer announced a temporary suspension of active development in November 2025, and the official documentation carries a notice that transfer or sale of the project is under consideration.
Key Features
- Three ways to display a character — VRM, Live2D, and PNG: VRM (3D models) supports gesture control through motion tags, and Live2D (Cubism 3 and later) as well as PNG characters that switch between expression variants are also supported
- Broad coverage of major LLMs: OpenAI, Anthropic, Google Gemini, Azure OpenAI, Groq, Cohere, Mistral AI, Perplexity, Fireworks, xAI, DeepSeek, OpenRouter, and Dify, plus local LLMs via LM Studio and Ollama
- A wide choice of speech synthesis engines: VOICEVOX, AivisSpeech, Aivis Cloud, Style-Bert-VITS2, Koeiromap, Google TTS, ElevenLabs, Cartesia, GSVI, and the OpenAI and Azure OpenAI TTS services, so free local engines and commercial cloud services can be mixed freely
- Automatic replies to YouTube Live comments: Comments are pulled through the YouTube API or OneComme, and the character reads them aloud and responds. This is the core feature for AITuber streaming
- WebSocket external linkage and slide mode: Speech content can be pushed in from an external server, and a presentation mode reads slides aloud automatically, extending the kit beyond streaming
- Advanced modes such as realtime API and screen recognition: Low-latency conversation via the OpenAI Realtime API, a reasoning mode that shows the thought process, automatic game commentary from screen capture, and person detection through facial recognition are included as more experimental features
Pricing
The kit itself is published as open source, but from v2.0.0 onward it adopts a custom license that treats non-commercial and commercial use differently.
| Plan | Price | Scope and conditions |
|---|---|---|
| Non-commercial use | Free | Personal non-profit use, educational institutions and educators, non-profit organization activities, demos at exhibitions and events |
| Standard license | ¥100,000 (one-time) | Use within one company, or redistribution to one client. Unlimited locations and users |
| Developer license | ¥300,000 (one-time) | Redistribution to up to five third parties. Unlimited locations and users per recipient |
| Enterprise license | ¥1,000,000 (one-time) | Unlimited redistribution, locations, and users |
| Custom license | Contact for details | Individual arrangements for requirements outside the tiers above |
All commercial licenses are one-time purchases with no expiration date. However, due to contractual constraints with Live2D Inc., commercial use of the Live2D features is not permitted (Live2D features remain available for non-commercial use). If a project starts out as non-commercial and commercial use is later confirmed, license fees are charged retroactively from the point commercial use began.
Pricing is current as of August 2026. Check the official license document for the latest terms. API usage fees for LLMs and speech synthesis are billed separately by each service provider.
Pros & Cons
✅ Pros
- Everything an AITuber stream needs — character display, speech synthesis, comment retrieval, LLM responses — is integrated from the start, so there is no need to assemble the pieces yourself
- The range of LLM and speech synthesis options is unusually wide, and cloud APIs and local execution can be swapped depending on the use case and budget
- A local LLM plus local speech synthesis setup runs without any metered external API charges
- Free for personal, educational, and non-profit use, which also makes it approachable as a learning subject
- The source code is public, so you can read how responses and streaming integration are actually put together
⚠️ Cons
- Active development was suspended in November 2025, so new features and bug fixes should not be expected at a steady pace
- The official documentation states that transfer or sale of the project is under consideration, leaving the future governance unsettled
- Commercial use requires purchasing a one-time license, so anyone monetizing a hobby project needs to check the cost and conditions first
- Live2D features carry an explicit restriction against commercial use
- Setting up a Node.js environment, obtaining API keys, and preparing character models are all required, making the barrier to entry higher than a no-code service
Comparison with Similar Services
| Criteria | AITuberKit | ChatVRM | VTube Studio | Character.AI |
|---|---|---|---|---|
| Provider | tegnike (individual developer, Japan) | pixiv | Denchi | Character Technologies |
| Main use | Building AI character chat and AITuber streaming | VRM character conversation demo | Live2D model tracking for streaming | Chatting with AI characters |
| Character models | VRM, Live2D, PNG | VRM | Live2D | Illustration images |
| Stream comment integration | YouTube supported | None | Works with streaming software (no AI responses) | None |
| Delivery | Self-hosted | Self-hosted | Desktop application | Web service |
| Pricing | Free for non-commercial / one-time commercial license | Open source | Free (some features paid) | Free plan / paid plan available |
Who Is It For
- People who want to start streaming as an AITuber but would rather not wire up comment retrieval and speech synthesis themselves
- Anyone who wants to give an existing VRM or Live2D character the ability to hold a conversation
- Those who want to run an AI character on a local LLM and local speech synthesis without worrying about metered API costs
- Developers who want to learn how an AI character app is structured by reading code that actually runs
- Teams that need a conversational character quickly for an exhibition demo or an educational purpose
Summary
AITuberKit gathers everything needed for AI character conversation and AITuber streaming into a single open-source kit out of Japan. Its breadth of LLM and speech synthesis options, support for local execution, and coverage of both VRM and Live2D give it a level of practical completeness that is rare in this space. On the other hand, active development is currently suspended and transfer of the project is under consideration, so anyone adopting it for long-term operation should first confirm they can maintain the code themselves. A realistic path is to run a local LLM setup within the non-commercial scope, then review the license terms before moving into monetization.