Back to Blog
Guide
4 min read
April 15, 2026

How to Use Sentry MCP with Claude Code

Search issues, inspect error details, manage releases, and triage bugs in Sentry directly from Claude Code — turning error monitoring into an integrated part of your development flow.

sentryerror-monitoringdebuggingobservabilityclaude-code

What is the Sentry MCP?

The Sentry MCP connects Claude Code to your Sentry organization, allowing it to search and read error issues, inspect stack traces, correlate errors with releases, and help you understand and fix production bugs faster. Claude becomes an intelligent triage assistant with full context from both your codebase and your error log.

Installation

mcpizy install sentry

Create a Sentry Auth Token at sentry.io/settings/account/api/auth-tokens/ with project:read and event:read scopes. Provide the token and your organization slug during setup.

Key Capabilities

  • Search and list issues — query issues by status, level, tag, or full-text search.
  • Inspect error details — read full stack traces, breadcrumbs, and request context for any event.
  • Correlate with releases — find which release introduced a regression.
  • Manage issue status — resolve, ignore, or merge issues programmatically.
  • Query performance data — inspect transaction performance and identify slow endpoints.

Example Usage

A new error is trending in production. Ask Claude to diagnose it:

// Claude will:
// 1. Fetch the full stack trace for the error
// 2. Read the breadcrumbs leading up to the crash
// 3. Open the relevant source file in your codebase
// 4. Identify the root cause and suggest a fix

Tips and Best Practices

  • Configure source maps in Sentry so Claude receives meaningful stack traces instead of minified code.
  • Use tags in your Sentry events (e.g., user_plan, region) so Claude can filter errors by business context.
  • Combine with the GitHub MCP to automatically open a GitHub issue for any Sentry error Claude cannot immediately fix.
  • Set up release tracking in Sentry — Claude can then compare error rates before and after a deploy.

View on MCPizy Marketplace

Found this useful? Share it.

MCP Servers Mentioned

SentryGrafanaGithub
All ArticlesBrowse Marketplace