npm install @llmrank.app/mcp

MCP Server for AI Coding Agents

27 SEO tools for Claude Code, Cursor, VS Code, Windsurf, and ChatGPT. Crawl sites, score pages across 37 AI-readiness factors, and fix issues — all from your IDE via the Model Context Protocol.

Quick Start

1. Generate an API token at Dashboard → Settings → API Tokens. 2. Add the MCP server to your IDE.

Claude Code

claude mcp add llm-boost \
  --env LLM_BOOST_API_TOKEN=llmb_xxx \
  -- npx -y @llmrank.app/mcp

Cursor

~/.cursor/mcp.json
{
  "mcpServers": {
    "llm-boost": {
      "command": "npx",
      "args": [
        "-y",
        "@llmrank.app/mcp"
      ],
      "env": {
        "LLM_BOOST_API_TOKEN": "llmb_xxx"
      }
    }
  }
}

Claude Desktop

claude_desktop_config.json
{
  "mcpServers": {
    "llm-boost": {
      "command": "npx",
      "args": [
        "-y",
        "@llmrank.app/mcp"
      ],
      "env": {
        "LLM_BOOST_API_TOKEN": "llmb_xxx"
      }
    }
  }
}

VS Code

.vscode/mcp.json
{
  "servers": {
    "llm-boost": {
      "command": "npx",
      "args": [
        "-y",
        "@llmrank.app/mcp"
      ],
      "env": {
        "LLM_BOOST_API_TOKEN": "llmb_xxx"
      }
    }
  }
}

Windsurf

~/.codeium/windsurf/mcp_config.json
{
  "mcpServers": {
    "llm-boost": {
      "command": "npx",
      "args": [
        "-y",
        "@llmrank.app/mcp"
      ],
      "env": {
        "LLM_BOOST_API_TOKEN": "llmb_xxx"
      }
    }
  }
}

ChatGPT / HTTP

Endpoint: https://mcp.llmrank.app/v1/mcp
Auth: OAuth 2.1 with PKCE (auto-discovered via /.well-known/oauth-authorization-server)

Environment variables: LLM_BOOST_API_TOKEN (required, starts with llmb_) and LLM_BOOST_API_URL (optional, defaults to https://api.llmrank.app).

27 Available Tools

Every tool includes MCP annotations (readOnlyHint, destructiveHint) so your agent knows which tools are safe to call without confirmation.

Projects

3 tools
  • list_projectsList all projects with domains and latest scores
  • get_projectGet project details including latest crawl score
  • create_projectCreate a new project by providing a domain

Crawling

3 tools
  • start_crawlStart a crawl scoring pages across 37 AI-readiness factors
  • get_crawl_statusCheck crawl progress in real time
  • list_crawlsGet crawl history with scores and timestamps

Scores & Analysis

5 tools
  • get_site_scoreOverall AI-readiness score with category breakdown
  • compare_scoresCompare two crawls to see improvements or regressions
  • get_score_historyScore trends over time across crawls
  • list_pagesList crawled pages with scores, sortable by grade
  • get_page_detailsFull page analysis: per-category scores, issues, fixes

Issues & Fixes

3 tools
  • list_issuesAll issues grouped by severity and category
  • get_fix_recommendationAI-generated fix steps with code examples
  • generate_fixGenerate code snippets to resolve a specific issue

AI Visibility

3 tools
  • check_visibilityCheck if your brand appears in AI search results across 6 platforms
  • list_visibility_historyTrack AI search presence over time
  • suggest_queriesAI-suggested queries to monitor based on your content

Content & Technical

4 tools
  • analyze_contentEvaluate content across 37 AI-readiness factors
  • suggest_meta_tagsGenerate optimized title, description, and OG tags
  • check_llms_txtValidate your llms.txt for AI crawler permissions
  • validate_schemaCheck structured data (JSON-LD, Schema.org)

Strategy

5 tools
  • get_recommendationsPrioritized action plan ranked by effort and impact
  • get_content_gapsTopics competitors cover that you don't
  • discover_keywordsAI-powered keyword discovery with search volume
  • list_competitorsCompetitors with AI-readiness score comparison
  • compare_competitorSide-by-side comparison with a specific competitor

Reports

1 tool
  • generate_reportComprehensive Markdown report with scores, issues, and recommendations

Pre-built Prompts

Common workflows packaged as MCP prompts your agent can invoke directly.

site-audit

Full AI-readiness audit: scores, critical issues, and prioritized action plan

fix-plan

Generate specific code and content fixes for your top issues

competitor-analysis

Compare your site against competitors and identify gaps

HTTP Transport & OAuth

For cloud-based agents like ChatGPT, the MCP server is available over Streamable HTTP with OAuth 2.1 authentication (PKCE, RFC 8414, RFC 9728).

MCP Endpoint

https://mcp.llmrank.app/v1/mcp

OAuth Discovery

/.well-known/oauth-authorization-server

Resource Metadata

/.well-known/oauth-protected-resource

Supports MCP Authorization Spec: Dynamic Client Registration (RFC 7591), PKCE with S256, WWW-Authenticate with resource_metadata, and token refresh.

Frequently Asked Questions

What is the Model Context Protocol (MCP)?

MCP is an open standard created by Anthropic that lets AI coding agents connect to external tools and data sources. It provides a unified way for IDEs like Claude Code, Cursor, VS Code, and ChatGPT to call tools — in this case, the 27 SEO tools from LLM Rank.

Which AI coding agents support MCP?

Claude Code, Cursor, Claude Desktop, VS Code (via Copilot), Windsurf, ChatGPT, and Perplexity all support MCP servers. LLM Rank works with all of them — use stdio transport for local IDEs or the HTTP endpoint for cloud-based agents.

Do I need a paid plan to use the MCP server?

The MCP server works with any LLM Rank plan, including the free tier. Free accounts can crawl up to 10 pages per crawl with 2 crawls per month. Paid plans unlock higher limits and additional features like competitor analysis.

Is the MCP server open source?

Yes, the @llmrank.app/mcp package is MIT licensed and published on npm. The source code is available on GitHub.

How does the HTTP transport work for ChatGPT?

ChatGPT and other cloud-based clients use the Streamable HTTP transport at mcp.llmrank.app/v1/mcp. Authentication uses OAuth 2.1 with PKCE — the client discovers the authorization server via the /.well-known/oauth-authorization-server endpoint and completes the flow automatically.

Start using LLM Rank from your IDE

Create a free account, generate an API token, and add the MCP server to your IDE in under 2 minutes.