moeru-ai / airi

2026-02-28 · Source: Github Trending: All languages · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Gaming & Interactive Media · Depth: Advanced, extended

Summary

Project AIRI aims to create open-source "cyber living souls" or virtual characters, inspired by Neuro-sama, that can interact, play games, and chat beyond simple text-based roleplaying. The project emphasizes ownership and accessibility, allowing users to have their digital companions available anytime, anywhere. AIRI is built using modern web technologies like WebGPU, WebAudio, and WebAssembly, while also supporting native NVIDIA CUDA and Apple Metal for desktop versions to ensure performance. It features capabilities such as memory systems, in-browser inference, speech recognition, speech synthesis via ElevenLabs, and support for VRM and Live2D models. The project is actively seeking developers, artists, and designers to contribute to its various sub-projects and expand its functionalities, including integrations with numerous LLM API providers like ChatGPT, Claude, and Ollama.

Key takeaway

For Machine Learning Engineers and AI developers interested in virtual character creation, Project AIRI offers a robust, open-source framework. You should consider contributing to its development, especially if you have expertise in WebGPU, computer vision, or reinforcement learning, to help advance the capabilities of interactive digital companions. The project's hybrid web and native approach provides a flexible foundation for building and deploying AI-driven virtual entities across various devices, enabling broader accessibility and functionality.

Key insights

AIRI creates open-source, interactive AI virtual characters using a hybrid web and native technology stack.

Principles

• Prioritize user ownership of digital life.
• Combine web and native technologies for broad accessibility.
• Foster community contribution for project growth.

Method

AIRI integrates large language models, memory systems, speech recognition, and synthesis with 2D/3D avatar rendering (VRM/Live2D) across web, desktop, and mobile platforms, leveraging WebGPU and native GPU acceleration.

In practice

• Explore AIRI for custom virtual companion development.
• Contribute to sub-projects like `unspeech` or `xsai`.
• Utilize AIRI's multi-platform deployment capabilities.

Topics

• AI Virtual Characters
• Large Language Models
• Multi-platform AI
• Virtual Character Interaction
• WebGPU

Code references

• moeru-ai/airi
• SillyTavern/SillyTavern
• huggingface/candle
• moeru-ai/chat
• moeru-ai/airi-factorio

Best for: Machine Learning Engineer, AI Engineer, Software Engineer, Creative Technologist

Related on AIssential

• OiiOii AI makes animation feel like directing, not prompt engineering — OiiOii AI streamlines animation production by managing the entire workflow, ensuring story continuity and reducing iterative costs.
• 5 AI Anime Waifu Generators You Should Try — AI anime waifu generators offer diverse features for character creation, interaction, and visual art, catering to various user preferences.
• Creating anthropomorphic animals with Gemini and HIggsfield — AI tools can merge static anthropomorphic images with human motion videos to create dynamic content.
• Google Vids gets AI upgrade with Veo and Lyria models, directable AI avatars — Google Vids integrates Veo 3.1 and Lyria AI models, plus directable avatars, for enhanced video and music creation.
• 10 Best Yodayo Alternatives in 2026: I Tested 10 Anime AI Art Platforms So You Don’t Have To — Yodayo alternatives address specific user frustrations like character consistency, cost, and multi-style generation.
• Testing Out the New Opus 4.8 👌 — Anthropic's Opus 4.8 demonstrates exceptional one-shot visual design capabilities for complex web interfaces.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Github Trending: All languages.