Scout Monitoring MCP
Scout APM data in Claude Code and Cursor: slow endpoints, traces and error groups with line numbers to fix code in place
- 30 ★ on the repository
- Last commit
Extended access
Why this level
- The API key opens all account monitoring data, including traces with request details
- Traces and errors may carry fragments of user data that the agent will see
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/scout-monitoring-mcpDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The API key opens all account monitoring data, including traces with request details
- Traces and errors may carry fragments of user data that the agent will see
This is third-party code. Review the repository files before installing.
What it does
The official local MCP server of the Scout Monitoring service. It reads application monitoring data through the Scout API and hands it to the agent: response time and throughput metrics, request traces and grouped errors. Eight tools cover the path from an app list down to a single trace. list_apps shows applications, get_app_endpoints gathers endpoints with aggregate metrics, get_app_trace returns a trace with all spans, get_app_error_groups lists fresh error groups, and get_app_insights finds N+1 queries, memory bloat and slow queries. Besides tools, the server exposes Scout setup templates for Rails, Django, Flask and FastAPI as resources. An agent in the editor matches a trace or error backtrace to code and suggests a fix, and can open an issue in GitHub, GitLab or Jira from the data.
Who it is for. Teams on Rails, Django, FastAPI and Laravel that already send data to Scout APM and want to work through slowdowns and errors from the editor.
Good fit when
- An endpoint got slower after a release and the agent needs its metrics and fresh traces
- An error keeps recurring in production and the backtrace needs a concrete code line to fix
- After a deploy you want to check for new N+1 queries
- A tracker ticket needs a trace and error frequency attached
- You are connecting Scout to a new Django or Rails app and want a ready config template
Not a fit when
- The app is not connected to Scout: the server reads only what the monitoring agent already sent
- You need host infrastructure logs or metrics: Scout MCP has none, Grafana or Datadog fit better
- There is no Docker or Python environment, and running without them is not covered
- You need to change monitoring settings: access is read-only
Example request
Find the slowest endpoints of my-app over 7 days and show average time, throughput and P95.MCP
- Transport
- stdio
- Authentication
- API key
Environment variables
SCOUT_API_KEYrequired, secret- Scout API key from the settings page, read-only
Limitations
The server will not start without a Scout API key, the one from the settings page, different from the agent key. The key is read-only but opens all account data.
The recommended run mode needs Docker. The authors say they deliberately do not trim response size, so it is better to raise the MCP output limit in Claude Code to 50,000 tokens.
Scout is a paid service, and an account plus the monitoring agent in the app are needed first. Availability of Scout from Russia was not checked.
How to disable. Remove the server with claude mcp remove scoutmcp and revoke the API key in Scout settings.
FAQ
How do I connect Scout MCP to Claude Code?
With claude mcp add scoutmcp, the SCOUT_API_KEY variable and a Docker run of the scoutapp/scout-mcp-local image. For a guided setup, npx @scout_apm/wizard configures Cursor, Claude Code or Claude Desktop.
Which key is needed, the agent key or the API key?
The API key from the Scout settings page. The monitoring agent key does not work. The API key gives read-only access to account data.
Which frameworks does Scout support?
The description names Rails, Django, FastAPI, Laravel and others. Setup templates in the server resources exist for Rails, Django, Flask and FastAPI.
What does the agent see in a trace?
The whole trace with every span and execution detail. For errors it gets the backtrace, from which the agent finds the line in the project code.
Can the hosted server replace Docker?
The MCP registry has a separate entry, com.scoutapm/scout-monitoring, for a hosted server at scoutapm.com/mcp. The README covered here describes only the local run and does not detail the hosted option.
Discussion
No questions or reviews yet.
Related
MCP serversMeasure MCP: an agent reads mobile app crashes, ANRs, traces, sessions and bug reports and looks for the root cause.
Datadog's official managed MCP server: logs, metrics, traces and incidents right in your agent
Sentry's official MCP server: search issues, inspect events and traces and analyze root causes from a coding agent
A single-container web dashboard for monitoring and controlling a fleet of 1C MCP servers: status, indexing progress, VRAM and license checks
The built-in OneUptime MCP: an agent works incidents, reads logs and traces, acknowledges alerts and sets up monitors.
Monitoring and status page service with an MCP server: an agent reads check state and manages incidents.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Scout Monitoring MCP (this card) | yes | API key | 30 | |
| Datadog MCP Server | yes | OAuth | 48 | broad observability platform with logs and infrastructure |
| Sentry MCP | no | API key | 961 | errors and releases only, without app performance traces |
| Grafana MCP Server | yes | API key | 3,663 | Grafana dashboards and metrics, while Scout covers APM |
README in short
The repository holds the local server code and a Node setup wizard. The image is published on Docker Hub, and releases go to PyPI and the MCP registry as com.scoutapm/scout-mcp-local. A separate registry entry covers the hosted version at scoutapm.com.