Open-source interface for interactive 3D character communication
Top 37.4% on sourcepulse
Amica provides an open-source interface for interactive communication with customizable 3D characters, enabling natural voice chat and emotional expression. It targets users interested in AI-driven virtual companions, offering a flexible platform for integrating various AI technologies for speech, vision, and language processing.
How It Works
Amica leverages a modular architecture, integrating technologies like three.js for 3D rendering and @pixiv/three-vrm for character display. It utilizes Transformers.js for in-browser AI model execution, Whisper for speech recognition, and Silero VAD for voice activity detection. For chatbot functionality, it supports multiple backends including Llama.cpp server, ChatGPT API, Ollama, and others, allowing users to choose their preferred LLM. Text-to-speech is handled by options like Eleven Labs API, OpenAI TTS, and local Coqui or RVC models.
Quick Start & Requirements
git clone git@github.com:semperai/amica.git
), then run npm install
.npm run dev
. Access at http://localhost:3000
.npm
). Configuration requires setting up API keys for chosen AI services (e.g., OpenRouter, Eleven Labs, OpenAI) in a .env.local
file.npm run tauri dev
for Tauri desktop development.Highlighted Details
Maintenance & Community
The project is actively maintained, with recent updates to version 1.2. Community engagement can be found via Twitter (https://twitter.com/arbius_ai) and GitHub issues.
Licensing & Compatibility
The majority of the project is released under the MIT license. Assets (3D models, images) are under their respective author licenses. MIT license generally permits commercial use and linking with closed-source projects.
Limitations & Caveats
Windows users are advised to create a new folder for Amica during installation to prevent potential file deletion during uninstallation. Documentation updates may lag behind feature releases.
1 week ago
1 week