MCP Index
Error MonitoringFreeactiveReviewed by MCPIndex

Sentry MCP

Official hosted Sentry MCP server for AI-driven error triage, issue search, and stack trace analysis inside your AI client.

View on GitHubView on NPMBrowse all tools

Looking for more MCP servers? Browse the full MCP tools directory or explore more tools in Error Monitoring.

OverviewConfigSetupFAQRelated

Quick overview

The average production incident burns 40+ minutes before a human even opens the correct stack trace — Sentry MCP collapses that window to seconds by inverting the debugging flow: instead of you pulling context out of the observability pipeline, the agent pulls distributed trace context, breadcrumbs, and suspect commits directly into its reasoning loop. The architectural insight most teams miss: a stack trace is a lossy projection of distributed causality, and Sentry is the only layer in the chain that reassembles it — release metadata, demangled source maps, async payload parsing across service boundaries — into something an LLM can actually reason about. Connected to Claude or Cursor, the server exposes tools like find_issues, find_events, and get_issue_details over an OAuth-secured remote transport, turning contextual AI debugging into a first-class workflow: paste an issue URL and the agent correlates the error fingerprint with the exact release and commit that introduced it. Self-hosted and EU data-residency instances are supported via host overrides, which matters far more than most setup guides admit.

What this MCP server is best for

  • Investigating production errors, stack traces, and release regressions through Sentry MCP when something breaks in a live application.
  • Useful for reviewing high-volume exceptions, identifying affected environments, tracing recent regressions, and understanding which issues need immediate attention.
  • Best for engineering teams that use Sentry for monitoring and want AI-assisted error triage, faster debugging, and clearer incident context.

When to choose it

Choose Sentry MCP when you want an MCP server focused on Error Monitoring and need tighter integration with your existing tools.

Good fit

sentryerror-monitoringobservabilitydebuggingstack-tracesapm

Sentry MCP Configuration

Use the following configuration as a starting point for Claude Desktop or any compatible MCP client, then replace placeholder credentials with your own values.

claude_desktop_config.json
{
  "mcpServers": {
    "sentry": {
      "url": "https://mcp.sentry.dev/mcp"
    }
  }
}

How to set up Sentry MCP

These setup steps cover the typical installation flow for this MCP server.

  1. 1

    Add the remote server URL https://mcp.sentry.dev/mcp to your client config — no local process is required because Sentry runs a fully hosted, OAuth-secured endpoint.

  2. 2

    Trigger the OAuth flow from your client (Claude Desktop prompts automatically on the first tool call) and authorize only the organization you actually debug — the consent screen scopes tokens per-org.

  3. 3

    Pro-Tip: If your Sentry org lives on EU data residency (eu.sentry.io) or self-hosted infrastructure, the default endpoint will authenticate successfully yet return an empty issue universe — set the host override (SENTRY_HOST env var, or the --host flag on the stdio build) BEFORE blaming the agent. This single misconfiguration causes the majority of Sentry MCP returns nothing reports.

  4. 4

    For clients without remote MCP support, bridge with stdio: npx -y mcp-remote https://mcp.sentry.dev/mcp — the bridge forwards OAuth headers transparently.

  5. 5

    Verify with one prompt: List my unresolved Sentry issues from the last 24 hours. If the agent returns issues with event counts and release tags, the pipeline is wired correctly.

  6. 6

    Restrict the Sentry auth token to project:read, event:read, and org:read scopes — never grant issue:write on shared workspaces unless you intentionally want an agent resolving and assigning issues autonomously.

Frequently asked questions

Common questions for Sentry MCP.

Why does Sentry MCP report no unresolved issues when my dashboard shows hundreds?

This is the classic silent failure: no error is thrown because authentication succeeded — but you are querying the wrong data plane. Sentry shards data by region (US vs EU) and by organization slug; a token issued against sentry.io cannot see eu.sentry.io events, and the server returns an empty list instead of a 401. The engineering fix: pin the correct host (SENTRY_HOST=https://eu.sentry.io or your self-hosted domain), explicitly pass the project slug in your prompt, and confirm the token org scope. If issues appear immediately after the host override, you were silently querying an empty region.

How does it handle minified JavaScript stack traces?

Sentry demangles frames server-side using your uploaded source maps before the agent ever sees them. If your release artifacts were uploaded with a mismatched dist value, the agent receives raw minified frames and will hallucinate file names — verify source map upload per release in your CI pipeline.

Can Sentry MCP resolve or assign issues autonomously?

Yes, but only if the token carries issue:write scope. Keep production tokens read-only and create a separate scoped token for write workflows — otherwise a vague prompt like clean up old errors can bulk-resolve issues you still need.

Does it work with self-hosted Sentry?

Yes. Use the stdio build with SENTRY_HOST pointing at your instance, or mcp-remote against your domain. The hosted endpoint only covers sentry.io SaaS organizations.

Sentry MCP vs Competitors

FeatureSentry MCPCompetitor
Root-cause context per error✅ Grouped issue with breadcrumbs, release tag, and suspect commit attached❌ Raw log lines; correlation is manual grep work across Kibana
Time from alert to suspect commit✅ Seconds — agent queries the issue fingerprint directly❌ Minutes to hours of query tuning and dashboard spelunking
Minified JS demangling✅ Automatic via uploaded source maps❌ Not available in log pipelines
Query interface✅ Natural language through the agent❌ KQL, Lucene, or CloudWatch Insights syntax required
Arbitrary unstructured log search at petabyte scale❌ Scoped to error events and transactions✅ Purpose-built for brute-force log retention and search

Related Guides

How to Install MCP Servers

Cross-client installation guide for Claude Desktop, Cursor, and VS Code.

Claude Desktop MCP Setup

Beginner-friendly setup guide for Claude Desktop MCP.

Best Open Source MCP Tools on GitHub

Compare popular open source MCP tools by use case and setup difficulty.