MCP 服务器概览

what is CompetLab MCP Server?

CompetLab MCP Server is a Model Context Protocol (MCP) server that exposes competitive intelligence data for B2B SaaS, with special focus on AI Visibility — how LLMs (ChatGPT, Claude, Gemini) mention and rank brands. It consolidates monitoring across five dimensions: Tech & Trust, Content, Positioning, Pricing, and AI Visibility, plus dashboards, historical trends, alerts, and synthesized strategic briefings.

how to use CompetLab MCP Server?

Remote (recommended): point any MCP-compatible client to https://mcp.competlab.com/mcp and authenticate with an API key via the CL-API-Key header or api_key query parameter.
- Example clients: Claude Code, Claude Web/Desktop, Cursor, VS Code (Copilot), Windsurf, Cline.
- Example (Claude Code): claude mcp add --transport http --header “CL-API-Key: YOUR_KEY” competlab https://mcp.competlab.com/mcp
Local (stdio): clone the repo, npm install && npm run build, then run with COMPETLAB_API_KEY=YOUR_KEY node dist/index.js. Useful for desktop clients that use stdio transport.
Get an API key: sign up at app.competlab.com/register, go to Organization Settings > API Keys (keys start with cl_live_).
Use tools by calling MCP methods (JSON-RPC over HTTP or stdio). Many tools are read-only; three async tools start scans and return scanId to poll.

key features

AI Visibility tracking: per-provider (OpenAI, Claude, Gemini) mention rates and visibility scores (0-100).
Five-dimension CI monitoring: Tech & Trust, Content, Positioning, Pricing, AI Visibility.
Dashboards, paginated history, per-run detailed reports, and trend endpoints (up to 200 data points).
Strategic briefings that synthesize changes and recommended actions.
Alerts and scheduling for automated monitoring.
33 MCP tools across 11 groups (projects, tech & trust, content, positioning, pricing, AI visibility, briefings/alerts, free public checks, async scans).
Free public tools: live sitemap checks, AI-crawler checks, fetch_url with JS rendering and optional cleanHtml, and async scans (tech stack, trust signals, agent adoption).

use cases

Monitor competitor pricing and detect changes to plans or billing models.
Track how LLMs recommend or rank your brand vs competitors (AI-driven go-to-market insights).
Run tech-stack and trust-signals scans on public domains to benchmark security and tech adoption.
Automate alerts for critical changes (content removals, security downgrades, AI-visibility drops).
Feed competitive data into an AI assistant (Claude, Copilot, Cursor) for on-demand briefings and strategic recommendations.

FAQ

How do I authenticate? Use CL-API-Key header for most clients; use ?api_key=YOUR_KEY for Claude Desktop/Web. For local stdio use COMPETLAB_API_KEY env var.
How do I get an API key? Sign up at app.competlab.com/register and create a key under Organization Settings > API Keys. Keys start with cl_live_.
Are there rate limits? fetch_url is limited to 60 requests/min per API key; other free tools have higher limits. Async scans return a scanId you poll until complete.
What if I see api_key_missing or api_key_invalid errors? Ensure the key is passed in the correct place (header, query param, or env var). Keys must start with cl_live_ and be the expected length.
Connection refused / timeout? Verify the server URL is exactly https://mcp.competlab.com/mcp (no trailing slash) and that your client supports the chosen transport (http or stdio).
Is it free? CompetLab offers a 14-day free trial (no credit card). MCP access is included with CompetLab subscriptions (documented pricing; platform is commercial).
Licensing? Repo documentation and configs are MIT; the platform/service is commercial. For support, use GitHub Issues or support@competlab.com.