Rampify MCP Server - Claude SEO Checker

Turn Claude Code into a powerful SEO checker. Real-time site audits, Google Search Console integration, and AI-powered recommendations directly in your editor (Cursor, Claude Code).

Get Started with Rampify →

What is a Claude SEO Checker?

A Claude SEO checker is an MCP (Model Context Protocol) server that adds SEO validation capabilities to Claude Code and Cursor. Unlike AI rank trackers that check if your site appears in Claude's AI responses, Rampify analyzes your website's technical SEO, validates meta tags, detects issues, and provides actionable fix recommendations—all from your terminal or IDE.

Not an AI rank tracker. This is a developer tool that brings SEO intelligence into your coding workflow.

Claude SEO Checker vs AI Rank Trackers

Keywords: claude seo checker, mcp seo server, seo tools for claude code, cursor seo tools, ai seo checker, claude code seo

Bring Google Search Console data, SEO insights, and AI-powered recommendations directly into your editor. No context switching, no delays.

Why Rampify?

• Real-time SEO intelligence in your editor (Cursor, Claude Code)

• Google Search Console integration - See clicks, impressions, rankings

• Content strategy insights - Discover what to write next based on real search data

• AI-powered recommendations - Fix issues with one command

• Pre-deployment checks - Catch SEO issues before they go live

• Zero context switching - Stay in your workflow

Learn more →

Installation

Prerequisites

• Node.js 18 or higher

• Rampify account (free to start)

Install via npm

The global installation makes the rampify-mcp command available system-wide.

Usage

Get Your API Key

Before configuring the MCP server, get your API key:

1. Sign up for Rampify (free to start)

2. Go to your Rampify dashboard

3. Navigate to Settings → API Keys

4. Click "Generate New Key"

5. Copy the key (starts with sk_live_...)

6. Use it in the configuration below

Quick Setup for a Project (Claude CLI)

Recommended: Configure MCP server per-project so each project knows its domain:

Now you can use MCP tools without specifying domain:

• get_page_seo - Automatically uses your project's domain

• get_issues - Automatically uses your project's domain

• crawl_site - Automatically uses your project's domain

Global Setup (Claude CLI)

For global access across all projects (must specify domain in each request):

Manual Configuration (Cursor)

Add to your Cursor settings UI or ~/.cursor/config.json:

Manual Configuration (Claude Code)

Add to your Claude Code MCP settings:

Configuration Options

Environment Variables

• BACKEND_API_URL (required): Rampify API endpoint - always use https://www.rampify.dev

• API_KEY (required): Your API key from Rampify dashboard (starts with sk_live_...)

• SEO_CLIENT_DOMAIN (optional): Default domain for this project (e.g., yoursite.com)

• CACHE_TTL (optional): Cache duration in seconds (default: 3600)

• LOG_LEVEL (optional): debug, info, warn, or error (default: info)

How to Use Tools

Discovering Available Tools

Ask Claude directly:

Claude will show you all available tools with descriptions.

Natural Language vs Direct Calls

Recommended: Use natural language (Claude will pick the right tool)

Alternative: Call tools directly (if you know the exact name)

Common Workflows

After Deployment:

Before Deployment:

Regular Monitoring:

Content Planning:

Available Tools

Quick Reference

1. get_page_seo

Get comprehensive SEO data and insights for a specific page. Works with both production sites AND local dev servers!

Parameters:

• domain (optional): Site domain (e.g., "example.com" or "localhost:3000"). Uses SEO_CLIENT_DOMAIN env var if not provided.

• url_path (optional): Page URL path (e.g., "/blog/post")

• file_path (optional): Local file path (will be resolved to URL)

• content (optional): Current file content

Examples:

Production Site:

Local Development Server:

Response includes:

• Source indicator: production_database, local_dev_server, or direct_content

• Fetched from: Exact URL that was analyzed

• Performance metrics (clicks, impressions, position, CTR) - only for production

• Top keywords ranking for this page - only for production

• Detected SEO issues with fixes - works for both local and production

• Quick win opportunities

• AI summary and recommendations

Local Development Workflow

Test pages BEFORE deployment:

1. Start your dev server:

   npm run dev # Usually runs on localhost:3000

2. Query local pages:

   Ask Claude: "Check SEO of localhost:3000/blog/draft-post"

3. Fix issues in your editor, then re-check:

   Ask Claude: "Re-check SEO for this page on localhost"

4. Deploy when clean!

What gets analyzed locally:

• Title tags

• Meta descriptions

• Heading structure (H1, H2, H3)

• Images and alt text

• Schema.org structured data

• Internal/external links

• Search performance (not available for local - GSC data only exists for production)

Response format:

2. get_issues

Get SEO issues for entire site with health score. Returns a comprehensive report of all detected problems.

Parameters:

• domain (optional): Site domain (uses SEO_CLIENT_DOMAIN if not provided)

• filters (optional):

  • severity: Array of severity levels to include (['critical', 'warning', 'info'])

  • issue_types: Array of specific issue types

  • limit: Max issues to return (1-100, default: 50)

Examples:

Response includes:

• Health score (0-100) and grade (A-F)

• Issue summary by severity (critical, warning, info)

• Detailed list of issues with fix recommendations

• Recommended actions prioritized by impact

Use cases:

• Site-wide SEO audits

• Finding all problems at once

• Tracking improvements over time

• Prioritizing fixes by severity

3. get_gsc_insights (NEW)

Get Google Search Console performance data with AI-powered content recommendations. Discover what to write next based on real search data.

Parameters:

• domain (optional): Site domain (uses SEO_CLIENT_DOMAIN if not provided)

• period (optional): Time period for analysis - 7d, 28d, or 90d (default: 28d)

• include_recommendations (optional): Include AI-powered content recommendations (default: true)

Examples:

What it provides:

1. Performance Summary

• Total clicks, impressions, average position, CTR

• Compare performance across time periods

2. Top Performing Pages

• Top 20 pages by clicks

• Each with performance metrics and top queries

• See what content resonates with your audience

3. Query Opportunities (4 types automatically detected)

• Improve CTR: High impressions (100+) but low CTR (<2%) → Optimize meta tags

• Improve Ranking: Position 6-20 → Push to page 1 with content improvements

• Keyword Cannibalization: Multiple pages competing for same query → Consolidate content

• Keyword Gap: High position (<5) but low volume → Expand content to target related queries

4. AI-Powered Content Recommendations

• High-priority topics based on search data

• Target queries for each recommendation

• Prioritized by potential impact (high/medium/low)

5. Query Clustering

• Groups related queries into topic themes

• Identifies topic authority opportunities

• Suggests comprehensive content pieces

Response includes:

Use cases:

Content Strategy:

Performance Optimization:

Keyword Research:

Topic Authority:

When to use:

• Weekly content planning sessions

• Quarterly content strategy reviews

• After publishing new content (check performance)

• When looking for low-hanging fruit (page 2 rankings)

• Before creating new content (avoid cannibalization)

Requirements:

• Google Search Console must be connected (connect in Rampify dashboard)

• Site must have some search traffic (impressions)

• GSC data synced (happens automatically weekly, or trigger manually)

Pro tips:

• Start with 28-day period for balanced view (not too recent, not too old)

• Use 7-day period to track recent changes

• Use 90-day period for seasonal trends

• Combine with generate_meta to optimize high-opportunity pages

• Run after GSC sync completes for latest data

4. crawl_site

Trigger a fresh site crawl and analysis. This is an active operation that fetches and analyzes all pages.

Parameters:

• domain (optional): Site domain (uses SEO_CLIENT_DOMAIN if not provided)

Examples:

What it does:

1. Discovers all URLs (via sitemap or navigation crawl)

2. Checks each URL (status, speed, SEO elements)

3. Detects issues (missing tags, errors, broken links)

4. Updates database with current state

5. Automatically clears cache so next get_issues or get_page_seo shows fresh data

Response includes:

• Total URLs found

• URLs checked

• Issues detected

• Crawl duration

• Crawl method (sitemap vs navigation)

When to use:

• After deploying code changes

• After fixing SEO issues

• Before running get_issues to ensure fresh data

• Weekly/monthly for monitoring

Note: This is the only tool that actively crawls your site. get_issues and get_page_seo just fetch existing data.

5. generate_schema

Auto-generate structured data (schema.org JSON-LD) for any page. Detects page type and generates appropriate schema with validation.

Parameters:

• domain (optional): Site domain (uses SEO_CLIENT_DOMAIN if not provided)

• url_path (required): Page URL path (e.g., "/blog/post")

• schema_type (optional): Specific schema type or "auto" to detect (default: "auto")

Supported schema types:

• Article / BlogPosting - Blog posts, articles, news

• Product - Product pages, e-commerce

• Organization - About pages, company info

• FAQPage - FAQ pages with Q&A

• BreadcrumbList - Auto-added for navigation

Examples:

What it does:

1. Fetches page HTML (local or production)

2. Analyzes content (title, description, author, date, images)

3. Detects page type from URL patterns and content

4. Generates appropriate JSON-LD schema

5. Validates schema and warns about placeholders

6. Returns ready-to-use code snippets

Response includes:

• Detected page type

• List of recommended schemas

• Generated JSON-LD for each schema

• Validation results with warnings

• Code snippets (Next.js or HTML)

• Implementation instructions

Use cases:

• Fixing "missing schema" warnings from get_issues

• Adding rich snippets for better search visibility

• Enabling Google Discover eligibility (requires Article schema)

• Improving CTR with enhanced search results

Example output:

Pro tip: After generating schema, test it with Google Rich Results Test

6. generate_meta (Enhanced with Client Profile Context)

Generate optimized meta tags (title, description, Open Graph tags) for a page. Now uses your client profile to generate highly personalized, business-aware meta tags that align with your target audience, brand voice, and competitive positioning.

Parameters:

• domain (optional): Site domain (uses SEO_CLIENT_DOMAIN if not provided)

• url_path (required): Page URL path (e.g., "/blog" or "/blog/post")

• include_og_tags (optional): Include Open Graph tags for social sharing (default: true)

• framework (optional): Framework format for code snippet - nextjs, html, astro, or remix (default: "nextjs")

NEW: Client Profile Integration

The tool automatically fetches your client profile and uses context like:

• Target keywords → Ensures they appear in title/description

• Target audience → Adjusts tone and technical depth

• Brand voice → Matches your preferred tone (conversational, technical, formal)

• Differentiators → Highlights unique selling points for better CTR

• Primary CTA → Ends description with appropriate call-to-action

Examples:

What it does:

1. Fetches page HTML (local or production)

2. Analyzes current meta tags (title, description)

3. Extracts content structure (headings, topics, word count)

4. Detects page type (homepage, blog_post, blog_index, product, about)

5. Identifies key topics from content

6. Returns analysis for AI to generate optimized meta tags

7. Provides framework-specific code snippets

Response includes:

• Page analysis:

  • Current title and description

  • Main heading and all headings

  • Word count and content preview

  • Detected page type

  • Key topics extracted from content

  • Images for OG tags

• Current issues:

  • Title too short/long

  • Meta description too short/long

  • Missing meta tags

• AI-generated meta tags:

  • Optimized title (50-60 characters)

  • Compelling meta description (150-160 characters)

  • Open Graph tags (if requested)

  • Twitter Card tags (if requested)

  • Ready-to-use code for your framework

Use cases:

• Fixing "title too short" or "description too short" warnings

• Improving click-through rate (CTR) from search results

• Optimizing social media sharing (OG tags)

• Aligning meta tags with actual page content

• A/B testing different meta descriptions

Real-World Impact: Before vs. After

Without Profile Context (Generic):

With Profile Context (Target audience: developers, Differentiators: "real-time collaboration, 50% faster"):

Profile Warnings System:

If your profile is incomplete, you'll get helpful warnings:

Or if no profile exists at all:

Example workflow:

Setting Up Your Profile:

To get the most value from generate_meta:

1. Visit /clients/{your-client-id}/profile in the dashboard

2. Fill out key fields:

   • Target Audience (e.g., "developers and technical founders")

   • Target Keywords (e.g., "real-time collaboration, dev tools")

   • Brand Voice (e.g., "technical but approachable")

   • Your Differentiators (e.g., "50% faster than competitors")

   • Primary CTA (e.g., "try_free" or "request_demo")

3. Use the tool - Profile context is automatically applied

4. See better results - Meta tags now match your business context

SEO Best Practices (Built-in):

• Title length: 50-60 characters (includes brand name if space allows)

• Description length: 150-160 characters (compelling call-to-action)

• Keyword placement: Primary keywords near the start

• Uniqueness: Each page gets unique meta tags based on its content

• Accuracy: Meta tags reflect actual page content (no clickbait)

Framework-specific output:

Next.js (App Router):

HTML:

Pro tips:

• Run after fixing content to ensure meta tags match

• Test social sharing with Facebook Sharing Debugger

• Monitor CTR improvements in Google Search Console

• Update meta tags when page content significantly changes

Development

Watch Mode

This will recompile TypeScript on every change.

Testing Locally

The MCP server will connect to your local backend at http://localhost:3000.

Debug Logging

Set LOG_LEVEL=debug in your .env file to see detailed logs:

Architecture

Caching

The MCP server caches responses for 1 hour (configurable via CACHE_TTL) to improve performance.

Cache is cleared automatically when:

• Entries expire (TTL reached)

• Server restarts

• You manually clear (not yet implemented)

Troubleshooting

"No client found for domain"

Solution: Add the site to your dashboard first at http://localhost:3000

"Backend API connection failed"

Checklist:

1. Is the backend running? (npm run dev in root directory)

2. Is BACKEND_API_URL correct in .env?

3. Check logs with LOG_LEVEL=debug

"MCP server not appearing in Cursor"

Checklist:

1. Did you build the server? (npm run build)

2. Is the path absolute in Cursor config?

3. Restart Cursor after changing config

4. Check Cursor logs (Help → Toggle Developer Tools → Console)

Empty or missing data

Common causes:

• Site not analyzed yet (run analysis in dashboard first)

• GSC not connected (connect in dashboard settings)

• No URLs in database (trigger site analysis)

"Could not connect to local dev server"

Solution:

1. Make sure your dev server is running (npm run dev)

2. Verify the port (default is 3000, but yours might be different)

3. Use full domain with port: localhost:3000 (not just localhost)

4. Check dev server logs for CORS or other errors

Example error:

Local vs Production Confusion

How to tell which source you're using:

Every response includes explicit source and fetched_from fields:

Pro tip: Set SEO_CLIENT_DOMAIN per project to avoid specifying domain every time:

• For local dev: SEO_CLIENT_DOMAIN=localhost:3000

• For production: SEO_CLIENT_DOMAIN=yoursite.com

Roadmap

Phase 1: Core Tools (Complete)

• DONE: get_page_seo - Get SEO data for a specific page

• DONE: get_issues - Get all site issues with health score

• DONE: get_gsc_insights - Get GSC performance data with content recommendations (NEW)

• DONE: generate_meta - AI-powered title and meta description generation

• DONE: generate_schema - Auto-generate structured data (Article, Product, etc.)

• DONE: crawl_site - Trigger fresh site crawl

Phase 2: Workflow & Optimization Tools (Planned)

• PLANNED: suggest_internal_links - Internal linking recommendations

• PLANNED: check_before_deploy - Pre-deployment SEO validation

• PLANNED: optimize_blog_post - Deep optimization for blog content

• PLANNED: optimize_landing_page - Conversion-focused SEO

Future (Phase 4+)

• Bulk operations across multiple pages

• Historical trend analysis

• Competitive monitoring

• Advanced AI insights and recommendations

Support

Need help?

• Documentation - Complete guides and tutorials

• GitHub Issues - Report bugs or request features

• Rampify Settings - Manage your sites and API keys

Learn More

• What is Rampify? - Product overview

• MCP Server Guide - Detailed documentation

• Blog - SEO tips and product updates

License

MIT