Vibeleaderboard
← Back to Tools
Index / tool

Sentry MCP

https://github.com/getsentry/sentry-mcp
Visit mcp.sentry.dev
Category
AI Agents
Rank
No. 341Tools index
Listed in
#9 Connect tools with MCP
Pricing
Open Source
Type
TOOL
Builder
getsentry
GitHub
845 stars
Latest release
0.39.0
Date

About

MCP server for Sentry. Lets Claude and other agents read issues, traces, and replays straight from your Sentry org.

What it does

Sentry MCP sits between a coding assistant and Sentry’s API. It exposes debugging and management operations as selectable tools, applies organization or project constraints, and returns structured results with text fallbacks for older clients.

Why it's ranked here

The project shows unusually serious production engineering for an agent integration: two transports, scoped tool exposure, multiple authentication paths, telemetry, unit tests, evaluations, and realistic API mocks. Its narrow focus on human-guided debugging also keeps the tool surface purposeful.

What's good

Remote OAuth and direct-token authentication cover different client capabilities. Operators can limit exposed skill groups, while server-side constraints keep requests tied to an organization or project. Structured responses retain compatibility with clients that only understand text. Self-hosted users can disable unsupported features.

Tradeoffs

The standard-input transport is explicitly described as unfinished. Natural-language search needs a separately configured model provider, although direct Sentry query syntax still works. Self-hosted installations may lack features such as Seer. This is intentionally not a complete interface to every Sentry capability.

How to use it well

Use it when developers already investigate Sentry data through Cursor, Claude Code, or another MCP client. Prefer the hosted service for Sentry SaaS and standard input for adapting self-hosted installations. Narrow skill access and project constraints. Keep broader Sentry administration outside this workflow.

Technical notes+

The pnpm TypeScript monorepo requires Node 22.13 or newer. packages/mcp-core/src/server.ts supports legacy and modern MCP SDK servers, filters tools by granted skills, injects request constraints, instruments execution, and converts structured-only results into compatible text payloads. packages/mcp-cloudflare/src/server/app.ts builds the Hono HTTP surface with OAuth metadata, CSRF checks, secure headers, chat and search routes, and a removed SSE endpoint returning 410. packages/mcp-server/src/index.ts implements the stdio CLI, device authentication, provider resolution, skill selection, telemetry, and graceful shutdown. packages/mcp-cloudflare/vitest.config.ts runs tests in workerd with mocked fetch and a test-only AJV compatibility alias.

Observed

License
FSL-1.1-ALv2
Primary language
TypeScript
Packaging
Private pnpm monorepo with a published @sentry/mcp-server package runnable through npx
Interfaces
Remote HTTP MCP service, stdio MCP transport, Claude Code plugin, and QA CLI
Platform support
Hosted Cloudflare service plus Node-based stdio support for Sentry SaaS and self-hosted installations
Testing
Unit tests, evaluations, manual testing playbooks, Cloudflare workerd tests, and MSW Sentry API mocks

Read from README.md, Makefile, package.json, packages/mcp-cloudflare/vite.config.ts, packages/mcp-cloudflare/vitest.config.ts, packages/mcp-cloudflare/worker-configuration.d.ts, packages/mcp-core/src/server.ts, packages/mcp-server/src/index.ts, packages/mcp-test-client/src/index.ts, packages/mcp-server-mocks/src/index.ts, packages/mcp-core/src/telem/index.ts, packages/mcp-core/src/tools/index.ts, packages/mcp-cloudflare/src/server/app.ts, packages/mcp-core/src/api-client/index.ts, packages/mcp-cloudflare/src/client/app.tsx.

What it can do

  • Read Sentry issues

    Sentry organization access → Issue data and details

  • Retrieve error traces

    Sentry organization access → Error trace information

  • Access replay sessions

    Sentry organization access → User session replay data

  • Query Sentry data through Claude

    Natural language queries about Sentry data → Structured responses about issues, traces, and replays

  • Enable AI agent analysis of Sentry data

    Sentry monitoring data → AI-powered insights and analysis

Tags

sentrymcpobservabilityagentintegration

Tech Stack

Node.js

Media

Sentry MCP

Comments (0)

No comments yet

Editorially curated, with community endorsements as a secondary signal. Corrections welcome.