diff --git a/content/docs/mcp-server.mdx b/content/docs/mcp-server.mdx index d1d0e68..3e28e23 100644 --- a/content/docs/mcp-server.mdx +++ b/content/docs/mcp-server.mdx @@ -25,8 +25,7 @@ npx @molecule-ai/mcp-server@1.0.0 "command": "npx", "args": ["@molecule-ai/mcp-server@1.0.0"], "env": { - "MOLECULE_URL": "http://localhost:8080", - "MOLECULE_API_KEY": "your-bearer-token" + "MOLECULE_URL": "http://localhost:8080" } } } @@ -37,13 +36,14 @@ npx @molecule-ai/mcp-server@1.0.0 **Pin the package version.** The examples above use `@1.0.0` — always specify an exact version and omit the `-y` flag. An unpinned `npx -y @molecule-ai/mcp-server` (no version) silently installs whatever npm serves on the next restart; if the package is ever compromised, it runs with your full MCP client permissions. Check [npm](https://www.npmjs.com/package/@molecule-ai/mcp-server) for the latest stable release before upgrading. -For SaaS deployments, set `MOLECULE_URL` to your tenant URL and add your API key: +For SaaS deployments, set `MOLECULE_URL` to your tenant URL: ```json -"MOLECULE_URL": "https://your-org.moleculesai.app", -"MOLECULE_API_KEY": "your-bearer-token" +"MOLECULE_URL": "https://your-org.moleculesai.app" ``` +Auth via `MOLECULE_API_KEY` is planned for a future release. + ### Verify Once configured, your MCP client should show 87 Molecule AI tools. Test with: @@ -154,34 +154,11 @@ The MCP server exposes tools across these categories: | Variable | Default | Description | |---|---|---| | `MOLECULE_URL` | `http://localhost:8080` | Platform API URL | -| `MOLECULE_API_KEY` | — | Bearer token for platform authentication | -| `MCP_SERVER_PORT` | `3000` | TCP port for HTTP/SSE transport | +| `MOLECULE_API_KEY` | — | Bearer token for platform authentication (planned) | -## Transport modes +## Transport -The MCP server supports two transport modes: - -### stdio (default, for local AI clients) - -Claude Code, Cursor, and similar local AI clients use stdio transport. The client spawns the MCP server as a child process and communicates via JSON-RPC over stdin/stdout. Configure in your AI client's `mcp.json` as shown in Quick start above. - -### HTTP/SSE (for remote agents and headless clients) - -For agents running on remote servers or in environments where stdio is not available, the MCP server exposes an HTTP endpoint with Server-Sent Events (SSE) for streaming responses. The agent card and tool responses are available at: - -``` -http://localhost:3000/agent/card # MCP server info + tool manifest -http://localhost:3000/ # SSE stream for responses -``` - -Start the server in HTTP/SSE mode: - -```bash -MCP_SERVER_PORT=3000 npm start -# Server starts on http://localhost:3000 -``` - -Configure your MCP client to connect over HTTP. The server emits `data: null` heartbeats every 30 seconds on idle connections to keep proxies and load balancers from closing the SSE stream. +The MCP server uses **stdio transport** — the client (Claude Code, Cursor, etc.) spawns the server as a child process and communicates via JSON-RPC over stdin/stdout. Configure in your AI client's `mcp.json` as shown in Quick start above. ## Troubleshooting @@ -189,5 +166,4 @@ Configure your MCP client to connect over HTTP. The server emits `data: null` he |---|---| | Connection refused | Check `MOLECULE_URL` points to running platform | | 401 Unauthorized | Token expired or revoked — create a new one | -| Port already in use | Set `MCP_SERVER_PORT` to an available port if 3000 is taken | | Tools not showing | Run `npx @molecule-ai/mcp-server@1.0.0` standalone to check errors |