A lightweight, local-first web application for chatting with multiple large-language models via OpenRouter. No signups, no accounts, just your conversation history, stored locally in your browser.

• Local-First: All conversation history is stored in your browser'''s localStorage. No data is stored on the server.
• Multi-Model: Switch between different LLM providers and models (OpenAI, Anthropic, Gemini, etc.) mid-conversation.
• Streaming Responses: See the AI'''s response in real-time as it'''s being generated.
• No Authentication: Start chatting immediately without needing an account.
• Minimalist UI: A clean, simple, and responsive chat interface.

• Frontend: React (Vite)
• Backend: Node.js (Express)

1. Clone the repository:

   git clone <repository-url>
   cd Compyle

    

2. Install backend dependencies:

   cd backend
   npm install

    

3. Install frontend dependencies:

   cd ../frontend
   npm install

    

4. Configure Environment Variables:

   • In the /backend directory, copy the .env.example file to a new file named .env.
   • Open the .env file and add your OpenRouter API key:

     OPENROUTER_API_KEY=your_secret_key_here
     

      

You need to run both the backend and frontend servers simultaneously in separate terminal windows.

1. Run the backend server:

   cd backend
   npm run dev

    

   The backend will typically run on http://localhost:3001.

2. Run the frontend server:

   cd frontend
   npm run dev

    

   The frontend will be available at http://localhost:5173 (or another port specified by Vite).

IMPORTANT: Your OPENROUTER_API_KEY is a secret and should never be exposed in the frontend code. This application is designed to keep the key secure on the backend. Do not commit your .env file to version control.