Sentry Mcp Server
Triage production issues, releases, and events from Sentry inside your coding agent instead of context-switching to the Sentry UI.
Overview
Sentry MCP Server is a MCP server for the Operate phase that connects AI assistants to Sentry issues, releases, events, and projects via the Sentry API.
What is this MCP server?
- stdio MCP server (sentry-mcp-server on npm) for Sentry.io API
- SENTRY_AUTH_TOKEN with org/project/event/issue read and issue write scopes
- Optional SENTRY_BASE_URL for self-hosted Sentry
- Surfaces projects, issues, releases, and events to the assistant
- Supports investigating and updating issues from chat during incidents
- Server version 1.0.0
- Default API base https://sentry.io/api/0/
- 5 documented auth token scopes
What problem does it solve?
When production breaks, copying Sentry URLs and issue JSON into chat burns time and loses release context the API already has.
Who is it for?
Indie operators who already use Sentry and want faster AI-assisted incident review and issue housekeeping.
Skip if: Projects with no Sentry project yet, or teams that forbid issue write access from automation.
What do I get? / Deliverables
With SENTRY_AUTH_TOKEN configured, your agent can query and act on Sentry issues and releases through MCP tools during triage and fixes.
- MCP tools backed by Sentry issues, events, releases, and projects
- Agent-assisted triage and issue updates against live Sentry data
- Optional self-hosted API targeting via SENTRY_BASE_URL
Recommended MCP Servers
Journey fit
Sentry is cataloged under Operate because its core job is live error tracking and issue response after you have shipped something worth monitoring. Errors is the natural shelf: issues, events, and release correlation are first-class incident and defect workflows, not greenfield feature build.
How it compares
Sentry API MCP integration, not an in-repo static analysis or test runner skill.
Common Questions / FAQ
Who is Sentry MCP Server for?
Builders and small teams using Sentry in production who want Claude Code, Cursor, or similar agents to read and update issues without leaving the IDE.
When should I use Sentry MCP Server?
Use it in operate/errors when triaging regressions, correlating releases with spikes, or drafting fixes backed by real Sentry event data.
How do I add Sentry MCP Server to my agent?
Register sentry-mcp-server as a stdio MCP server, set SENTRY_AUTH_TOKEN with the documented scopes, and optionally SENTRY_BASE_URL for self-hosted Sentry.