Sentry MCP Server

Sentry’s official MCP Server bridges AI agents and Sentry’s error tracking platform. Instead of navigating Sentry’s dashboard, ask your AI assistant to search errors, analyze stack traces, or trigger root cause detection directly from your IDE.

What It Does

The MCP server exposes Sentry’s API through standardized MCP tools. Developers can query production errors, retrieve stack traces, and initiate analysis without leaving their coding environment. Authentication uses OAuth (cloud) or tokens (self-hosted), reducing the need for manual API key management in development workflows. See Sentry MCP Docs for setup details.

Installation

Cloud Sentry:

claude mcp add --transport http sentry https://mcp.sentry.dev/mcp

Authenticate with your Sentry organization via OAuth. See Sentry MCP Setup for Claude Code, Cursor, and other tools.

Self-Hosted:

npx @sentry/mcp-server@latest --access-token=<token> --sentry-host=https://your-sentry.com

See Sentry MCP GitHub for self-hosted setup and authentication requirements.

Core Capabilities

• Search and retrieve error issues from projects
• Access full stack traces and event metadata
• Trigger Sentry’s analysis features
• Query organization and project configuration
• Track releases and deployments
• Create issues and manage alerts

Prerequisites

Cloud:

• Active Sentry account
• OAuth support in your AI tool

Self-Hosted:

• Sentry instance (documented versions supported)
• User Auth Token with appropriate scopes
• Network access to your Sentry instance

Use Cases

Error Investigation: Ask Claude Code to summarize critical errors from the last 6 hours, including affected users and stack traces.

Regression Detection: Compare error rates before and after deployment to identify new issues introduced in a release.

Automated Triage: Schedule an agent to periodically query unresolved errors and create GitHub issues with investigation context.

Limitations

• Search functionality may require external LLM provider configuration
• Token-based auth (self-hosted) requires manual credential management
• Sentry API rate limits apply
• Some advanced filtering options limited by MCP tool structure

Compared to Direct API

The MCP server provides a curated, AI-focused interface to Sentry’s API. Direct API access offers more granular control but requires custom integration code. The MCP approach exposes Sentry functionality via MCP tools for use in IDE-based AI assistants. See Sentry MCP GitHub for implementation details.

Community & Status

Sentry officially maintains this server. It works with Claude Code, Cursor, Claude Desktop, VS Code, and other MCP-compatible tools. Source code and documentation available at GitHub and Sentry Docs.