parsehex / ai-chat

ai-chat

Setup

Windows: Install Git For Windows (install the bash terminal)

1. npm install
2. In backend/, rename .env.sample to .env and fill in the values
3. npm run build
4. npm start

TTS

The app is setup to support ElevenLabs voices with an API key. SayTTS is used if supported on your platform (if so, saytts voices will show up in the returned available voices). I believe Mac and Windows is supported.

Alternatively, support for Coqui TTS is being added. Install with pip install TTS, get a list of models with tts-server --list_models, and start the server with e.g. tts-server --model_name tts_models/en/sam/tacotron-DDC.

When this app starts, it will request localhost:5002/voices and if there's a voice returned, the voice will show as an available voice in the app.

ai-chat's Issues

Data dirs not made on first run

Specifically, data/threads doesn't seem to be created (dataRoot folder + others are created), so when you try to create a thread you get in the server log:
ENOENT: no such file or directory, scandir '/home/user/ai-chat/data/threads'

Workaround is to create data/threads manually, then everything works fine.