Give your AI agent the same tools you use

Maple's MCP server lets agents query your traces, logs and metrics, using the same queries as the dashboard.

The surface

One endpoint, any MCP client

Connect Claude Code or any MCP client. It sees the same data you do.

/mcp api.maple.dev/mcp

In motion

Five steps from list_services to propose_fix, each one a query you could run yourself.

mcp · agent session LIVE

> ask maple

list services that errored in the last 15m

→ list_services { window: "15m", filter: "errors > 0" }

← 3 services

paymentserrors142p992.4s recoserrors38p99840ms shippingerrors6p99320ms

> investigate payments

→ find_errors { service: "payments", window: "15m" }

← 1 issue

stripe.charge declined · 142 events · 98 users

→ error_detail { issue_id: "iss_8a91" }

← root cause: idempotency key collision

sample trace: trace_9f3c · order 8421

→ propose_fix { issue: "iss_8a91", fix: "scope idempotency_key by user_id" }

← fix attached. opened pr#412.

4 tool calls · 4.2s open mcp protocol

What it does

Check a service: diagnose_service reports latency, error rate and throughput, and what changed.
Read the traces: find_slow_traces and inspect_trace return real spans, so the agent sees the actual queries.
Read your code: Link a repository and the agent can go from a stack trace to the exact line.
Changes need approval: Actions that change things are separate tools. In Maple's chat, they wait for your approval.
Limited to your org: Every query is locked to the org on the key. An agent can't see another team's data.