🔭 Raygun MCP Server

Investigate errors and performance in Raygun with AI assistants

A remote Model Context Protocol (MCP) server that connects AI assistants to Raygun Crash Reporting, Real User Monitoring, and Application Performance Monitoring.

📘 Setup Guide • 📚 Tool Reference • 🚀 Quick Start

---

✨ Key Features

• 🐛 Error Management - Investigate, resolve, and track application errors and crashes with full stack traces and context
• 🚀 Deployment Tracking - Monitor releases and correlate errors with deployments to identify problematic changes
• ⚡ Performance Insights - Analyze page load times, user metrics, and performance trends over time
• 🔬 APM Investigation - Find backend hotspots, recurring performance issues, and slow traces
• 👥 User Monitoring - Track customer sessions, behavior patterns, and identify affected users
• 🤝 Team Collaboration - Manage invitations and coordinate error resolution across your team
• 📊 Metrics & Analytics - Time-series analysis and distribution histograms for errors and performance

📋 Requirements

• A Raygun account with an active subscription
• An MCP client with OAuth support, or a Raygun Personal Access Token

🚀 Getting Started

The Raygun MCP server is hosted at:

https://api.raygun.com/v3/mcp

Raygun hosts and updates the service, so there is no package or local server to install.

OAuth (recommended)

Add the URL as a Streamable HTTP server without an Authorization header. On first connection, your MCP client will open Raygun in your browser so you can sign in, select a plan, review the requested permissions, and authorize access.

You do not need to supply an OAuth client ID or client secret. Raygun supports automatic client registration.

codex mcp add raygun --url https://api.raygun.com/v3/mcp
codex mcp login raygun

This configuration is shared by the Codex CLI, Codex IDE extension, and ChatGPT desktop app on the same Codex host.

In the ChatGPT desktop app, open Settings > MCP servers, select Add server, choose Streamable HTTP, and enter the Raygun endpoint. Save and restart the app, then select Authenticate for Raygun.

ChatGPT on the web uses remote MCP tools through Plugins and does not read your local Codex MCP configuration. A separately published Raygun Plugin would be installed from the Plugins catalog.

claude mcp add --transport http raygun https://api.raygun.com/v3/mcp

Run /mcp in Claude Code and follow the browser authentication flow.

{
  "mcpServers": {
    "raygun": {
      "url": "https://api.raygun.com/v3/mcp"
    }
  }
}

Cursor will prompt you to authenticate when it connects.

{
  "servers": {
    "raygun": {
      "type": "http",
      "url": "https://api.raygun.com/v3/mcp"
    }
  }
}

VS Code will prompt you to authenticate when it connects.

Personal Access Token

PAT authentication remains supported for clients without MCP OAuth support and for automation. The following example uses Cursor's configuration format; other clients use the same Authorization header in their HTTP server configuration:

{
  "mcpServers": {
    "raygun": {
      "url": "https://api.raygun.com/v3/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_PAT_TOKEN"
      }
    }
  }
}

💬 Your First Prompt

After configuration, try asking your AI assistant:

• “Find the errors that first appeared after the latest deployment of my checkout application.”
• “Investigate the most frequently occurring production error and show me a representative stack trace.”
• “Which backend methods or database queries consumed the most time in the last 24 hours?”
• “Show me the last session for customer user@example.com and explain where their experience failed.”
• “Compare p95 page performance over the last seven days and identify the largest regression.”

🛠️ Tools

The Raygun MCP server provides the following categories of tools:

• applications_list - List all applications in your Raygun account
• applications_search - Search for applications by name
• application_get_details - Get detailed application information
• application_regenerate_api_key - Generate a new API key for an application

• error_groups_list - List error groups within an application
• error_groups_search - Search occurrences by error, stack trace, URL, version, user, tags, or custom data
• error_group_investigate - Get complete details about a specific error group
• error_group_update_status - Change error group status (resolve, ignore, activate)
• error_group_add_comment - Add investigation notes to an error group
• error_group_read_comments - Review existing investigation notes
• error_group_instances_list - List occurrences in an error group
• error_instance_get - Get the full stack trace and context for an occurrence

• deployments_list - List deployments for an application
• deployment_create - Create a new deployment record
• deployment_get_latest - Get the most recent deployment with error analysis
• deployment_investigate - Get comprehensive deployment information
• deployment_manage - Update or delete a deployment

• customers_search - Search customers by name, email, or identifier
• customer_investigate - Get customer profile, recent error groups, and sessions
• sessions_list - List user sessions with environment and device data
• session_investigate - Inspect a session's environment, errors, and page views

• apm_issues_search - Find recurring backend performance issues and anti-patterns
• apm_issue_investigate - Inspect an issue's metrics and affected traces
• apm_traces_search - Find slow or failing backend requests
• apm_trace_investigate - Inspect profiler threads and significant calls in a trace
• apm_hotspots_search - Rank slow methods, database queries, and external calls

• pages_list - List monitored pages in an application
• page_investigate - Get page details for metrics queries
• metrics_website_performance_analyze - Track performance trends over time
• metrics_performance_distribution_analyze - Understand performance variability
• metrics_error_trends_analyze - Track error rates and patterns

• invitations_manage - List and review team invitations
• invitation_send - Send a new team invitation
• invitation_revoke - Cancel a pending invitation

For detailed documentation on each tool, see the Tool Reference.

🔐 Authentication

OAuth

Raygun implements the MCP authorization flow for HTTP clients. The server advertises its authorization server through OAuth Protected Resource Metadata, and compatible clients register themselves automatically using Client ID Metadata Documents or Dynamic Client Registration.

Access is tied to the Raygun user, plan, and scopes approved on the consent screen. Compatible clients securely store and refresh their own tokens.

Each OAuth connection grants access to one Raygun plan. Add another Raygun MCP configuration and authorize it separately when you need to work with a different plan.

Security and access

• OAuth never gives the client more access than the signed-in Raygun user already has.
• Read and write access is controlled by the scopes approved during OAuth or assigned to the PAT.
• Some tools can change Raygun data, including error statuses, deployments, invitations, comments, and ingestion API keys. Review write actions before approving them in your MCP client.
• Application tools can return or rotate ingestion API keys. Treat returned keys as credentials; rotating a key immediately invalidates the previous one.
• Tool results may be included in conversations processed by your chosen AI client or model provider. Follow your organization's policies for sharing diagnostic and customer data.

Obtaining a Personal Access Token

PATs remain available when OAuth is not suitable:

1. Navigate to Raygun Personal Access Tokens
2. Click Create New Token
3. Give your token a descriptive name (e.g., "MCP Server Access")
4. Select the appropriate permissions for your use case
5. Copy the generated token and send it as Authorization: Bearer YOUR_PAT_TOKEN

Keep PATs secure and never commit them to version control.

For more details, see the Raygun API documentation.

---

📖 About

The Raygun MCP server enables AI assistants to investigate Crash Reporting, RUM, and APM data; track deployments; analyze performance; and manage application monitoring through natural-language conversations.

🔗 Resources

• 📚 Raygun Documentation
• 📘 Raygun MCP Setup Guide
• 🔌 Raygun API Reference
• 🤖 Model Context Protocol
• 🐛 Report Issues

---

Built with ❤️ by Raygun