M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
L

LogClaw MCP Server

Official
Connect AI coding tools to LogClaw incidents, logs, and anomaly detections
How to install ↓ Compare GitHub source repository ↗
Category
Dev Tools #176 of 438
Stars
★ 79 Popular
Transport
stdio (local process)
Runtime
Node.js · Hosted (no local runtime)
Credentials
API key / credential required
Last commit
⚠
56FMRS · C

Official LogClaw MCP server exposing 8 tools over stdio so AI coding tools can access incidents, logs, and anomaly data. The source states the tool count but does not name the individual tools, so the tool list is left empty. A solid fit for teams already on LogClaw.

Strongest · Documentation 14/20 Weakest · Reliability 7/20

Reliability
7/20
Security and permissions
12/20
Maintenance
11/20
Documentation
14/20
Setup experience
12/20
Why each score
Reliability 7/20
The manifest claims 8 tools but neither the manifest nor the README enumerates their names, parameters, or example calls, so declared behavior cannot be checked against implementation. The README shows a plausible local dev/build path (npm install && npm run build && LOGCLAW_API_KEY=lc_proj_test npx .), but no CI workflow or committed test evidence for this specific MCP server was provided, so the init/tool-list handshake and error handling cannot be confirmed as reproducible. Per the static calibration rule, without verifiable CI+test evidence the score is capped well below the mid-range.
Security and permissions 12/20
Auth relies on a required LOGCLAW_API_KEY env var; the optional LOGCLAW_ENDPOINT defaults to an external vendor-hosted proxy (https://ticket.logclaw.ai), meaning by default log/incident-related traffic flows to a third-party managed service. This is disclosed rather than covert, but no least-privilege detail is given for the 8 tools (read vs. write scope, confirmation for any destructive actions), so the permission boundary is unverifiable. No red-line evidence (hardcoded secrets, credential theft, unconfirmed destructive defaults) was found, but incomplete scoping disclosure caps the score at a middle tier.
Maintenance 11/20
The listing metadata states license unknown, while the README badge and text both claim Apache 2.0 — an unresolved discrepancy that is deducted. Governance files (CONTRIBUTING.md, CODE_OF_CONDUCT.md, SECURITY.md) exist and the npm package is at version 1.0.4, indicating some release iteration; 79 stars and 24 open issues suggest visible activity, but no verifiable recent commit/release timeline was provided to assess response cadence, warranting a mid-range score.
Documentation 14/20
The project has a layered documentation set — a detailed README plus references to a dedicated MCP integration guide, API key docs, and architecture docs. However, for the MCP server itself, the README does not detail the 8 tools' parameters, output formats, rate limits, cost, or troubleshooting; that detail is deferred to external doc pages not included in the evidence, leaving a documentation gap for the component under review.
Setup experience 12/20
Setup is straightforward on paper: `npx logclaw-mcp-server` plus a single required env var, with dev build steps also provided. However, no mainstream-client configuration example (e.g., an mcpServers JSON snippet for Claude Desktop/Cursor) was included in the supplied material, and no CI or test evidence confirms a working connection in practice. Per the static calibration rule, absent verifiable execution evidence the score cannot exceed 15, so a constrained mid-to-upper score is given.

Static review · not runListed 2026-09-07

Read the FMRS scoring method →

Fit and risk

What it can accessUses the network

Best for

  • SRE and development teams already running LogClaw or using its cloud service
  • Engineers who want incident and log data directly in Claude Code, Cursor, or Windsurf

Not for

  • Users who do not use the LogClaw platform
  • Anyone without a valid LOGCLAW_API_KEY (lc_proj_ prefix)

Required permissions

  • Requires a valid LogClaw project API key (lc_proj_ prefix)
  • Network access to the LogClaw auth proxy endpoint (default https://ticket.logclaw.ai)

Risks and side effects

  • API key leakage could expose project incident and log data; secure environment variables accordingly
  • Logs exposed to AI tools may contain sensitive operational information
  • Depends on remote endpoint availability; the self-hosted local stack runs single-tenant with AI RCA disabled by default

Setup

Before you start

Runtime：Node.js · Hosted (no local runtime)

LOGCLAW_API_KEY requiredsecret Required LogClaw project API key starting with lc_proj_, obtainable from the LogClaw cloud console.
Other optional settings (1)
LOGCLAW_ENDPOINT optional Optional LogClaw auth proxy endpoint, defaults to https://ticket.logclaw.ai.
  1. Run via npx logclaw-mcp-server, or from the repo's apps/logclaw-mcp-server directory run npm install && npm run build and start it.
  2. Set the required environment variable LOGCLAW_API_KEY (a LogClaw project API key starting with lc_proj_).
  3. Optionally set LOGCLAW_ENDPOINT to your LogClaw auth proxy endpoint (default https://ticket.logclaw.ai).
  4. Configure the server in your MCP client using the stdio transport.

Check that it works

Run npx logclaw-mcp-server and confirm the 8 LogClaw MCP tools appear in your client's tool list (Claude Code, Cursor, Windsurf); then ask about recent incidents to verify the connection.

Troubleshooting

  1. Verify LOGCLAW_API_KEY is set and starts with lc_proj_
  2. If self-hosting or using a custom endpoint, check LOGCLAW_ENDPOINT points to the correct auth proxy
  3. Configure the client with stdio transport and inspect its logs
  4. Confirm the LogClaw backend services are reachable

Things to try

Once connected, you can ask your AI assistant things like:

  • Show me the most recent LogClaw incidents and their severity
  • Summarize recent anomalies detected for the payment-api service
  • List all unresolved anomaly detections in LogClaw
  • Show the root cause analysis for the latest incident

Use cases

Let AI coding assistants query active incidents and their severities in LogClaw
Retrieve and analyze logs and anomaly detections inside your IDE
Bring incident context into code-fixing workflows

Supported clients

Claude Code
Cursor
Windsurf

Listed from the project's documentation, not tested by this site.

Overview

The LogClaw MCP Server is the official MCP server for LogClaw, a log intelligence platform offering OTLP-native log ingestion, anomaly detection, trace correlation, and AI root cause analysis. It runs over stdio, is published on npm as logclaw-mcp-server (version 1.0.4), and exposes 8 tools that let MCP-compatible clients such as Claude Code, Cursor, and Windsurf query LogClaw incidents, logs, and anomaly detections directly. A hosted remote version is also available at https://mcp.logclaw.ai (OAuth 2.1).

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 63.3k · Tools 2 Compare with this →
Compare with other MCP servers

Source revision ff82088bd23a Data synced 2026-10-19 Read the FMRS scoring method