heyalice.app — template of a custom back-end in node.js

This repository is currently experimental and may undergo significant changes. The plan is to maintain full compatibility with the OpenAI interface.

Installation

1. Clone this repository
2. Install dependencies with bun bun install
3. Copy .env.example as .env and update your OpenAI API key
4. Start server using bun start or bun server.ts

Connection with "Alice" app

1. Open Settings by either: pressing Command + , or clicking on Search Bar and then Settings, or by clicking on cog icon (fullscreen mode)
2. Open Advanced tab
3. Add endpoint, by default it's http://localhost:3005/api/chat
4. (optional but recommended) Add authorization header
5. Add model you want to use on your back-end. It can't be in conflict with gpt-4o or any other model. It's a good practice to use prefix like for example — local:gpt-4o
6. From the list of the models (right beneath text input on the chat) select local:gpt-4o model or any model you set in 5th step.

Known issues

1. When server is down and you send the message, you need to reload the app (right mouse button -> reload) or restart it.