Raygun MCP

Raygun handles crash reporting, real user monitoring and application performance monitoring. Its MCP server bridges the Raygun API to an MCP client, so a coding assistant can pull the same error and session data a developer would otherwise open the Raygun dashboard to read. The server exposes error groups for an application along with the detail of a single group, including stack traces and the users affected, plus status changes on a group so an agent can resolve, ignore or reactivate it. Deployment records are readable with the commits attached to them, which is what ties an error spike back to a specific release. Customer and session listings are also available, down to the event trail of one session. Authentication uses a personal access token scoped to the Raygun account. The practical value is closing the gap between a stack trace and a candidate fix without copy and paste between two windows. The agent reads the failing frame, opens the matching source file in the repository, and proposes a patch. It is equally useful for triage after a release: sweep the new error groups, rank them by how many users hit them, resolve the ones already fixed and leave the rest annotated for a human.