Sentry MCP

Overview

Sentry MCP is a Model Context Protocol server that enables AI agents to interact with Sentry, the leading application monitoring and error tracking platform. It allows agents to fetch error reports, analyze issues, and help diagnose problems in production applications.

Key Features

🐛 Error Analysis

Fetch Issues: Retrieve error issues from Sentry projects
Issue Details: Get detailed information about specific errors
Error Trends: Analyze error frequency and patterns
Stack Trace Analysis: Understand error root causes

📊 Project Insights

List Projects: View all Sentry projects
Project Stats: Get error statistics per project
Release Tracking: Monitor errors by release version
Environment Filtering: Filter by environment (production, staging, etc.)

🔍 Search & Filter

Search Issues: Find specific errors by message or fingerprint
Filter by Tags: Filter issues by custom tags
Date Range: Query errors within specific time periods
Severity Levels: Filter by error severity

Installation

# Using npx (recommended)
npx -y @modelcontextprotocol/server-sentry

# With Sentry DSN
SENTRY_DSN=https://xxx@xxx.ingest.sentry.io/xxx npx -y @modelcontextprotocol/server-sentry

Configuration

Required Setup

Get a Sentry API Key:
- Log into Sentry
- Go to Settings → API Keys
- Create a new API key with appropriate permissions
- Recommended scopes: project:read, event:read

Set Environment Variables:

export SENTRY_AUTH_TOKEN=your-auth-token
export SENTRY_ORG=your-organization-slug

Configure in Claude Desktop:

{
  "mcpServers": {
    "sentry": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-sentry"],
      "env": {
        "SENTRY_AUTH_TOKEN": "your-auth-token",
        "SENTRY_ORG": "your-org-slug"
      }
    }
  }
}

Available Tools

Tool	Description
`list_projects`	List all projects in the organization
`get_issue`	Get details of a specific issue by ID or URL
`list_issues`	List issues with filtering options
`get_issue_stats`	Get error statistics for a project
`search_issues`	Search for issues by query
`get_releases`	List releases for a project
`get_tags`	List available tags for filtering

Usage Examples

List Your Projects

Show me all my Sentry projects

Returns:

Projects:
- my-app-web (web)
- my-app-api (api)
- my-app-mobile (mobile)

Get Issue Details

Show me details for issue ABC-123

Returns:

Issue: ABC-123
Title: TypeError: Cannot read property 'map' of undefined
Project: my-app-web
Status: Unresolved
First Seen: 2026-05-10
Last Seen: 2026-05-14
Occurrences: 247
Users Affected: 42

Search for Recent Errors

Find all errors in my-app-api from the last 24 hours

Agent queries Sentry with:

Project: my-app-api
Time range: Last 24 hours
Level: error

Analyze Error Patterns

What are the most common errors in production this week?

Agent returns:

TypeError: undefined is not a function - 1,234 occurrences
NetworkError: Failed to fetch - 892 occurrences
ValidationError: Invalid input - 567 occurrences

Get Error Statistics

Show me error trends for my-app-web over the last month

Returns daily error counts with trend analysis.

Filter by Tags

Show me errors tagged with 'critical' in the last week

Agent filters by tag: level:error tag:critical

Integration with AI Agents

Proactive Error Monitoring

Claude: "I noticed you have several unresolved errors in Sentry..."
[Lists recent critical issues]
Claude: "Let me analyze the most frequent one..."
[Fetches and analyzes stack trace]
Claude: "This appears to be caused by a null reference in the user profile component. Here's a potential fix..."

Incident Response

User: "We're seeing errors in production"
Claude: "Let me check Sentry for recent issues..."
[Queries Sentry for last hour]
Claude: "I found 3 new critical issues. Let me get details..."
[Provides analysis and potential fixes]

Release Monitoring

Claude: "Your latest release v2.3.1 has 47 new errors..."
[Lists new errors by release]
Claude: "Most are related to the new authentication flow. Here's what I found..."

Custom Agent Integration

from mcp import ClientSession, StdioServerParameters
from mcp.client.stdio import stdio_client
import os

# Connect to Sentry MCP
server_params = StdioServerParameters(
    command="npx",
    args=["-y", "@modelcontextprotocol/server-sentry"],
    env={
        "SENTRY_AUTH_TOKEN": os.environ["SENTRY_AUTH_TOKEN"],
        "SENTRY_ORG": os.environ["SENTRY_ORG"]
    }
)

async with stdio_client(server_params) as (read, write):
    async with ClientSession(read, write) as session:
        await session.initialize()
        
        # List projects
        projects = await session.call_tool("list_projects", arguments={})
        
        # Get recent issues
        issues = await session.call_tool(
            "list_issues",
            arguments={
                "project": "my-app-web",
                "status": "unresolved",
                "limit": 10
            }
        )
        
        # Get specific issue
        issue = await session.call_tool(
            "get_issue",
            arguments={"issue_id": "ABC-123"}
        )

Sentry API Permissions

Required API key scopes:

Scope	Required For
`project:read`	List projects, get issue details
`event:read`	Read error events and stack traces
`org:read`	Access organization settings
`release:read`	View release information

Pros

✅ Production Insights: Direct access to real error data
✅ Root Cause Analysis: AI can analyze stack traces
✅ Trend Detection: Identify error patterns over time
✅ Release Tracking: Monitor errors by version
✅ Proactive Alerts: Agents can notify about new issues
✅ Stack Trace Understanding: AI explains complex errors

Cons

❌ API Key Required: Must create Sentry API key
❌ Organization Setup: Requires Sentry organization
❌ Rate Limits: Sentry API has rate limits
❌ Cost: Sentry has costs for high-volume projects
❌ Read-Only: Cannot create issues or resolve them

When to Use

Choose Sentry MCP when:

You use Sentry for error tracking
You want AI to help diagnose production errors
You need to analyze error patterns
You want proactive error monitoring

Consider alternatives when:

You use a different error tracking service
You only need occasional error checks
You want write access to Sentry (not supported)

Resources

GitHub: https://github.com/modelcontextprotocol/servers/tree/main/src/sentry
Sentry Documentation: https://docs.sentry.io/
Sentry API: https://docs.sentry.io/api/
MCP Servers Repo: https://github.com/modelcontextprotocol/servers

Comparison

Feature	Sentry MCP	Logtail MCP	Datadog MCP
Error Tracking	✅ Excellent	⚠️ Basic	✅ Good
Stack Traces	✅ Full	⚠️ Limited	✅ Full
Trend Analysis	✅ Yes	⚠️ Basic	✅ Excellent
Release Tracking	✅ Yes	❌ No	✅ Yes
Setup Complexity	⚠️ Moderate	✅ Easy	⚠️ Moderate
Cost	⚠️ Can be high	✅ Low	⚠️ Can be high

Last updated: May 2026

Sentry MCP

Overview

Setup

Configuration

Documentation

Sentry MCP

Overview

Key Features

🐛 Error Analysis

📊 Project Insights

🔍 Search & Filter

Installation

Configuration

Required Setup

Available Tools

Usage Examples

List Your Projects

Get Issue Details

Search for Recent Errors

Analyze Error Patterns

Get Error Statistics

Filter by Tags

Integration with AI Agents

Proactive Error Monitoring

Incident Response

Release Monitoring

Custom Agent Integration

Sentry API Permissions

Pros

Cons

When to Use

Resources

Comparison