| # 🌐 GLMPilot Server | |
| The `server` package is the core backend for the GLMPilot IDE. It handles real-time coordination, acts as a secure proxy to the GLM Mini AI, and manages containerized terminal sessions. | |
| ## Key Responsibilities | |
| 1. **API Proxy**: Securely routes requests to the GLM Mini API without exposing the `GLM_API_KEY` to the client. | |
| 2. **WebSocket Management**: Uses `Socket.io` to provide real-time connection channels for the IDE. These channels include: | |
| - **Terminal Events**: Streams `stdout` and `stderr` to the client's xterm.js UI, while receiving `stdin`. | |
| - **Code Sync**: Relays editor state and file system changes between the client, the backend file system, and the AI context. | |
| 3. **Multi-Agent Reviewers**: Houses the logic for specialized AI agents (Security, Performance, Style, Documentation) situated in `src/agents/`. These agents can be triggered to analyze workspace code and summarize issues or generate auto-fixes. | |
| ## Main Entry Point | |
| - `src/index.ts`: Straps Express, Socket.io, and the API routes together. | |
| ## Development | |
| Run `npm run dev` (or `npm run dev:server` from the root) to start the server with hot-reloading using `ts-node-dev` or `nodemon`. | |