Built for AI Code
Chat with your app data in human language — Plug your AI into Scout
Scout MCP exposes performance issues and slow endpoints to your coding assistants. Available as a hosted server with zero setup, or self-hosted if you prefer to run it locally.
and more
Scout MonitoringBased on your monitoring data, your slowest endpoint is POST /api/users/profile/upload with an average response time of 3.2 seconds and P95 latency of 8.1 seconds. This endpoint handles user profile image uploads and is bottlenecked by synchronous image processing operations.
Would you like for me to come up with optimization ideas for it?
Scout MCP targets the unpredictable elements of AI coding
Lost flow
Context-switching kills momentum, and UX suffers
Tool sprawl
APM + errors + logs = separate, heavy, expensive
Black Box AI Code
Bugs after deploy that are hard to explain or reproduce
Makes AI Monitoring Natural
Scout MCP tracks any issues in the editor, automatically giving your agent deep insights
Ask your agent about app issues in natural language and get performance answers
Skip dashboards and get clear performance monitoring answers on demand.
Awareness-based performance fixes
With insights from Scout MCP, your agents learn to suggest contextual fixes for any detected issues.
Get Started
Add Scout MCP to your AI editor
Claude (web and desktop)
Scout Monitoring is available in Claude's connector directory. Open the listing, click Connect, and sign in to Scout. No setup needed.
Add to ClaudeHosted
Zero setup. Connect directly in Claude Code with one command. Authenticates via OAuth.
claude mcp add --transport http scout https://scoutapm.com/mcpRun the MCP server locally. Use the install wizard to set up everything you need.
npx @scout_apm/wizardKeep production running smoothly
Discover hidden issues
Surface AI-specific bugs before they hit production
Hassle-free monitoring
Errors, Logs, and Traces — right where you need them
Everything in the editor
Focus on informed delivery, no context-switching needed
FAQ
Scout MCP FAQ
What AI Assistants are supported?
Scout MCP works with any AI assistant that supports the Model Context Protocol, including Claude Desktop, Claude Code, Cursor, VS Code with Copilot, Windsurf, Cline/Roo, Amp, and more.
Is there a remote MCP server?
Yes. The hosted Scout MCP server is available with no infrastructure setup required. It authenticates via OAuth and works with any MCP-compatible AI assistant. In Claude, you can connect it straight from Claude's connector directory. Set-up docs here.
Does using Scout MCP have any associated cost?
There is no direct cost for using Scout MCP. It uses API access from your existing Scout account. Rate limits may apply, and there may be indirect token costs from your AI assistant provider.
What data does the Scout MCP server have access to?
The Scout MCP server has read-only access to insight, error, metric, and trace data from your Scout account. Log data support is coming soon.
Can I see what my AI coding agent itself is doing, not just what it can access?
That's a separate, early-stage feature we're building called agent monitoring. Scout MCP gives your agent access to Scout's data. Agent monitoring is the reverse: visibility into what the agent did. It's now in beta, and you can sign up here.