Agents
Connecting an agent
Connect Claude Code, Codex, Cursor, Claude Desktop or another MCP client to Mike Trace.
Mike Trace connects to AI agents over the Model Context Protocol (MCP), the open standard many agents already use to work with external tools.
Supported clients
- Claude Code
- Codex
- Cursor
- Claude Desktop
- Other MCP-compatible clients
How connecting works
- Add Mike Trace to your agent. Add Mike Trace as an MCP server in your agent’s settings. The exact endpoint and setup instructions for your account are shown in the Mike Trace Console.
- Sign in and approve. Your agent opens a browser page where you sign in to Mike Trace and approve the agent for a specific project. Clients that support standard OAuth sign-in handle this automatically.
- Start investigating. The agent can now read and investigate cases in that project.
You never paste your own password or session into a terminal or agent configuration.
What agents can access
Connected agents get scoped, expiring access limited to investigation work in the project you approved. They cannot see other projects or accounts, change your account settings, or authorize risky actions on your behalf.
Managing agents
From the Console you can see which agents are connected, rename them, disconnect them or replace their access at any time. Disconnecting an agent doesn’t erase its history: everything it recorded stays attributed to it.
Step-by-step setup guides for each client are available to customers in the Mike Trace Console.