Google Maps Extractor MCP

AI-powered lead generation from Google Maps. Search businesses, enrich with emails/phones/socials, score leads, and export to CSV -- all from your AI agent.

Not just a scraper. A complete lead gen pipeline that replaces $200/month tools like Apollo and Hunter.io.

Why This Tool?

Feature	This Tool	Apollo.io	Hunter.io	Outscraper
Free to use	Yes	$49/mo	$49/mo	Pay-per-row
MCP server (for AI agents)	Yes	No	No	No
Email discovery	Yes	Yes	Yes	No
Social links	Yes	Limited	No	No
Tech stack detection	Yes	No	No	No
Lead scoring	Yes	Yes	No	No
Google Maps search	Yes	No	No	Yes
Cross-platform (SERP + LinkedIn)	Yes	Partial	No	No
CSV/JSON export	Yes	Yes	Yes	Yes
No API key needed	Yes	No	No	No

How It Works

Google Maps search
  --> Detail extraction (phone, website, address, rating)
    --> Website enrichment (emails, socials, tech stack)
      --> SERP ranking check (optional)
        --> LinkedIn company lookup (optional)
          --> Lead scoring (0-100)
            --> Export CSV/JSON

One command to your AI agent:

"Find me 20 dental clinics in Dubai with their emails and phone numbers, export as CSV"

Quick Start

As MCP Server (Claude Desktop, Cursor, etc.)

Add to your MCP client config:

{
  "mcpServers": {
    "google-maps": {
      "command": "python",
      "args": ["/path/to/mcp-servers/google-maps/server.py"]
    }
  }
}

No API keys needed. Install dependencies:

pip install "mcp[cli]" playwright playwright-stealth
playwright install chromium

With Docker

docker build -t google-maps-mcp .
docker run google-maps-mcp

Tools

search_businesses

Quick Google Maps search. Returns name, rating, address, category.

{"query": "restaurants", "location": "Dubai Marina", "max_results": 20}

find_leads (the main tool)

Search + enrich + score. Returns emails, phones, socials, tech stack, lead score.

{"query": "real estate agencies", "location": "Miami", "max_results": 20, "enrich": true}

Example output:

{
  "leads": [
    {
      "name": "Sunshine Realty Group",
      "lead_score": 85,
      "rating": 4.8,
      "review_count": 127,
      "phone": "+1 (305) 555-0123",
      "website": "https://sunshinerealty.com",
      "emails": ["[email protected]", "[email protected]"],
      "social_links": {
        "linkedin": "https://linkedin.com/company/sunshine-realty",
        "instagram": "https://instagram.com/sunshinerealty",
        "facebook": "https://facebook.com/SunshineRealtyMiami"
      },
      "tech_stack": ["WordPress", "Google Analytics", "HubSpot"],
      "address": "1234 Brickell Ave, Miami, FL 33131",
      "maps_url": "https://maps.google.com/..."
    }
  ],
  "count": 20,
  "avg_score": 62.4,
  "enriched": true
}

generate_lead_report

Full pipeline with optional SERP ranking + LinkedIn company data.

{
  "query": "plumbers",
  "location": "London",
  "max_results": 15,
  "check_seo": true,
  "check_linkedin": false
}

get_business_details

Deep-dive on a single business.

{"maps_url": "https://maps.google.com/maps/place/..."}

get_reviews

Customer reviews for sentiment analysis.

{"maps_url": "...", "max_reviews": 50, "sort_by": "newest"}

export_leads

Export last results to CSV or JSON file.

{"format": "csv"}

CSV columns: lead_score, name, category, rating, review_count, phone, email, website, address, linkedin, facebook, instagram, twitter, tech_stack

Lead Scoring

Every lead gets a score from 0-100:

Signal	Points
Personal email (john.doe@...)	25
Any email	15
Phone number	15
Website	10
Social media links (5 each)	max 15
Rating >= 4.0	10
50+ reviews	10
Has address	5

Leads are sorted by score -- highest quality first.

Website Enrichment

When enrich=true, the tool visits each business website and extracts:

• Emails: From homepage + /contact + /about pages. Ranked by quality (personal > role-based > generic). Filters out noreply@, postmaster@, etc.
• Social links: LinkedIn, Facebook, Instagram, Twitter/X, YouTube, TikTok
• Phone numbers: From tel: links and visible text
• Tech stack: WordPress, Shopify, Wix, React, HubSpot, Intercom, Zendesk, Google Analytics, and more

Results are cached for 24 hours to avoid re-scraping.

Cross-Server Integration

When sibling MCP servers from the BlackHole suite are installed:

• SERP Scraper: Checks if leads rank on Google for the search query
• LinkedIn Scraper: Pulls company data (size, industry, specialties)

Both are optional -- works standalone without them.

Performance

• Browser geolocation auto-matched to 30+ cities worldwide
• Concurrent enrichment (3 parallel websites)
• Retry with exponential backoff on failures
• 24-hour result cache to avoid redundant scraping
• Headless browser with stealth anti-detection

License

MIT