CompNav · Open-source projects
Project AIRI
An open-source digital-life project that brings AI characters to browsers and desktops, combining text and voice interaction, Live2D/VRM avatars, and game or desktop capabilities.
What it is
AIRI is not just a chat window; it aims to place an AI character within the user's digital environment. It offers web and desktop versions for text and real-time voice interaction. The desktop app puts more emphasis on an avatar stage and computer-use capabilities, and the project showcases interaction with games such as Minecraft and Factorio. AIRI is evolving quickly, so feature maturity differs across platforms.
Core mechanisms and capabilities
The project has both web and desktop versions: the browser version is a way to try character interaction, while desktop adds a persistent character window and deeper system integration. Official clients are available for multiple platforms, though features and installation differ by platform.
Desktop characters can use formats such as Live2D, VRM, and MMD, bringing conversation, voice, and an avatar stage into one experience. Avatar assets may carry separate licenses, so AIRI's code license does not cover every model or resource.
AIRI's direction also includes voice conversation, desktop tools, and game interaction. These capabilities depend on the model, operating system, permissions, and current release; showcased scenarios should not be read as full support on every platform.
Editorial take
AIRI merits inclusion because it combines AI companionship, avatar presentation, and desktop interaction in an actively developed open-source project rather than offering only a model API or developer component. It has substantial community attention, and recent releases continue to expand voice, avatar, and desktop capabilities. Readers should still treat it as a fast-moving digital-character platform, not a finished desktop pet; they also need to choose a model provider or configure a local model.
Who it suits
Suited to people interested in self-hosted AI characters, desktop avatars, or AI VTuber workflows who are willing to choose a model, adjust settings, and work with a rapidly changing project.
Deployment and technical requirements
AIRI provides a browser experience and downloadable desktop app. The web version opens from the official site; the desktop app is distributed as installers with a separate setup manual. Having a client does not mean a free general-purpose model is bundled: model-provider setup is still required for first use.
Runtime requirements
Hosted models generally require signing in or providing an API key; local models require a compatible model service and model files. Desktop voice, avatars, and computer use also involve audio devices, system permissions, and additional resources. The project does not provide one universal hardware threshold for every feature.
First use requires selecting a chat model and provider: users can sign in to AIRI's official service, connect a hosted provider, or use a local service such as Ollama. Hosted options usually require an account or API key and may incur fees; local inference requires model downloads and compatible hardware. Desktop local inference is better suited to users with an appropriate GPU or Apple Silicon device; exact needs depend on model size.
Limitations and considerations
AIRI is actively developed and changes quickly, so feature availability can shift between versions. Hosted models involve provider costs and data transfer; local models reduce cloud dependence but require model files and compute resources. Third-party avatar and voice assets may use separate licenses.
Official sources and further reading
Information checked: Oct 26, 2026.