Connect your AI agent to Lampwick over MCP
Lampwick's MCP server lets Claude Code, Cursor, Codex, claude.ai and ChatGPT read your issues, requests, logs and monitors, and act on them if you allow it.
The MCP server at https://app.lampwick.dev/mcp gives an AI agent the same data you see in the dashboard. Every tool is an API operation, with the access you choose when you connect: Read-only (recommended) or Read & write.
What you get
- Issues, requests, queries, logs, uptime monitors and scheduled tasks, as tools named after their API operation (
listIssues,getIssue, …). - With Read & write: resolve issues and change monitors and projects. Never: delete a team, rotate an ingest token or change a browser key.
- Sign-in without a token in claude.ai, Claude Desktop, ChatGPT, Cursor, VS Code and Claude Code; a personal
lwp_…token for other clients.
Set it up
In Claude Code, add the server, then run /mcp and sign in from there:
claude mcp add --transport http lampwick https://app.lampwick.dev/mcp
In claude.ai, ChatGPT, Cursor or VS Code, add a custom connector with the URL https://app.lampwick.dev/mcp and leave any client ID and secret fields empty. Lampwick opens in your browser: log in, choose the access, and click Connect.
Ask "Which Lampwick projects can you see?" to check it works.
Ask your agent
Once it is connected, paste this in your app's repository:
Look at the open Lampwick issues of this project that were first seen since the last deploy. Pick the one with the most events, find its cause in this repository and propose a fix. Don't resolve the issue until I have reviewed the fix.
What an agent reads includes error messages and logs from your apps, which could contain instructions meant for it. With read-only access, those can't change anything.
Next step
The MCP docs have the token setup for Claude Code, Codex, Cursor, VS Code, Gemini CLI, Zed, Cline and the Claude and OpenAI APIs. No account yet? Your agent can set up Lampwick for you.