Amica
An open source AI character interface you can run locally, with voice chat, VRM avatars, and support for Ollama, llama.cpp, and cloud APIs.
Amica is a locally runnable interface for talking with customizable 3D AI characters. It's for people who want an animated, voiced character as the face of their AI assistant, with a choice of local LLM backends or cloud services. The project builds on Pixiv's ChatVRM.
You can import VRM character files and adjust the voice to suit the avatar. An emotion engine lets characters express feelings, and responses can include emotional expressions. Voice chat combines speech recognition with synthesized speech; vision support uses Bakllava.
Amica connects to Ollama, llama.cpp servers, KoboldCpp, and Oobabooga. It also works with LM Studio through a ChatGPT-compatible API. Those choices let you use a model running on your own hardware, while OpenRouter and the ChatGPT API provide cloud alternatives. Running the interface locally doesn't make a connected cloud service local: those services process the requests you send them.
Speech has similar flexibility. Whisper handles speech recognition, while voice output supports local Coqui alongside Speech T5, RVC, and AllTalkTTS. Eleven Labs and OpenAI are available as API options. Transformers.js supports running models in the browser.
The interface can run as a local web app, and Amica also has a Tauri desktop application. A hosted demo is accessible on mobile, tablet, and desktop. The code is open source under the MIT license; character models and images retain their creators' separate licenses.
