Logclaw Mcp Server
Wire your agent into LogClaw so it can pull live incidents, logs, and anomaly signals while you debug production.
Overview
LogClaw MCP Server is an Operate-phase MCP server that lets coding agents read LogClaw incidents, logs, and anomaly detections over stdio.
What is this MCP server?
- Stdio MCP server (npm logclaw-mcp-server v1.0.4) talking to LogClaw via an auth proxy
- Tools for incidents, logs, and anomaly detections from your LogClaw project
- Configured with LOGCLAW_API_KEY (lc_proj_*) and optional LOGCLAW_ENDPOINT
- Keeps investigation inside Claude Code, Cursor, or Codex instead of tab-hopping dashboards
- MCP integration layer, not a hosted skill—register the server in your agent config
- Package version 1.0.4 on npm as logclaw-mcp-server with stdio transport
- Default auth proxy endpoint https://ticket.logclaw.ai unless LOGCLAW_ENDPOINT overrides
Community signal: 79 GitHub stars.
What problem does it solve?
Production issues force you out of the editor to chase incidents and log lines in a separate observability UI.
Who is it for?
Indie builders already on LogClaw who want MCP-driven triage during post-launch debugging.
Skip if: Teams without LogClaw, or anyone who only needs local dev logs with no hosted incident store.
What do I get? / Deliverables
Your agent can fetch LogClaw incident and log context in-session so fixes start from real telemetry instead of guesses.
- Agent-callable access to LogClaw incidents, logs, and anomaly tools
- Documented env vars LOGCLAW_API_KEY and LOGCLAW_ENDPOINT for your workspace
Recommended MCP Servers
Journey fit
LogClaw is an external observability backend; the MCP bridge is used after ship when something breaks or drifts in prod. Incidents, log streams, and anomaly detections map directly to ongoing visibility and alert triage, not one-off launch tasks.
How it compares
Live observability MCP bridge to LogClaw, not a generic log-tail skill or self-hosted Loki setup.
Common Questions / FAQ
Who is LogClaw MCP Server for?
Solo and indie developers using LogClaw who want Claude Code, Cursor, or Codex to query incidents and anomalies without leaving the repo.
When should I use LogClaw MCP Server?
Use it in Operate when you are investigating outages, error bursts, or anomalies and need agent-accessible LogClaw data.
How do I add LogClaw MCP Server to my agent?
Install the npm package logclaw-mcp-server, set LOGCLAW_API_KEY and optionally LOGCLAW_ENDPOINT, then register the stdio server in your MCP client configuration.