Sentry MCP

The Sentry MCP Server connects AI assistants directly to your Sentry error tracking account, bringing production error context into the same conversation where you write and debug code. Instead of context-switching between your editor and the Sentry dashboard when hunting down a bug, you can ask your AI assistant what errors are currently happening in production and have it analyze them alongside your codebase.

Practical workflows this enables: ask your AI assistant to show the top 5 unresolved errors in your production environment, paste the stack trace into the chat for analysis, and request a fix suggestion — all without leaving your editor. The assistant can access real error data, see which errors are occurring most frequently, filter by release version, and read the full exception context including breadcrumbs and user attributes.

The server also surfaces regression detection: after deploying a new version, you can ask whether any new error signatures appeared and compare error rates to the previous release. This makes post-deploy validation faster and more systematic.

For on-call engineers, the Sentry MCP Server significantly reduces the time from alert to investigation. A natural language description of the error, combined with the ability to query related issues and see recent deployment history, compresses the context-building phase of incident response from minutes to seconds.

Installation requires a Sentry authentication token with the appropriate scopes. The server is open-source, maintained by Sentry, and compatible with any MCP client including Claude Desktop, Cursor, Windsurf, and Zed. It works with both Sentry SaaS and self-hosted Sentry instances.