Sentry MCP

Sentry MCP is the official Model Context Protocol server for Sentry, the error monitoring and performance platform. It exposes an organization's Sentry data to a coding agent: the teams and projects under an account, the issues those projects have recorded, and the individual events behind an issue with their stack traces, breadcrumbs, tags, and release context. It runs as a hosted remote server with OAuth, and can also be run locally against an access token for self-hosted installations. Beyond reading, the server covers the actions that follow an investigation. An agent can search issues in plain language rather than Sentry's query syntax, open a specific issue to see where in the code it originates, mark it resolved or ignored, assign it, and create projects or client keys when wiring up a new service. Sentry's automated root cause analysis can be invoked through the same connection, so the agent receives a proposed explanation and fix to evaluate against the repository it is working in. The point is closing the gap between an error in production and the line that caused it. A developer can ask which errors spiked after a release, have the agent read the stack trace, locate the function in the local checkout, propose a patch, and resolve the issue once it ships. It fits teams already running Sentry who want their coding assistant to see production failures instead of guessing at them.