Lead Gen MCP Server

Lead Gen MCP Server


A free MCP server that runs a full B2B outbound pipeline inside a Claude conversation.

Type one sentence. Get scraped leads → Google Sheet → Instantly campaign, all without leaving your chat.

You:    Scrape 30 HVAC companies in Austin TX, create a campaign called
        "HVAC Austin | Free Audit", and push the leads.

Claude: Done.
        → 28 qualified leads saved: https://docs.google.com/spreadsheets/d/...
        → Campaign created (ID: a1b2c3)
        → 28 leads uploaded. Ready to launch in Instantly.

What it does

| Tool | Description | |---|---| | scrape_leads | Google Maps → qualified leads → Google Sheet (via Apify) | | read_sheet | Load leads from any existing Google Sheet | | create_instantly_campaign | Create a new Instantly email campaign, returns campaign ID | | push_to_instantly | Upload leads array to an Instantly campaign |

Prerequisites

• Python 3.11+
• Apify account + API token (for Google Maps scraping)
• Instantly account + API v2 key (for email campaigns)
• Google Cloud project with OAuth credentials (for Google Sheets)

Install

git clone https://github.com/Sourav333444/lead-gen-mcp
cd lead-gen-mcp

python -m venv .venv
.venv\Scripts\activate        # Windows
# source .venv/bin/activate   # Mac/Linux

pip install "mcp[cli]" apify-client gspread google-auth google-auth-oauthlib requests python-dotenv

Configure

Create a .env file in the project root:

APIFY_API_TOKEN=apify_api_...
INSTANTLY_API_KEY=your_instantly_v2_key

Google Sheets auth (one-time)

Download OAuth credentials from Google Cloud Console → APIs & Services → Credentials → Create OAuth 2.0 Client ID (Desktop app). Save as credentials.json in the project root.

Then run:

python -c "
from google_auth_oauthlib.flow import InstalledAppFlow
flow = InstalledAppFlow.from_client_secrets_file('credentials.json', [
    'https://www.googleapis.com/auth/spreadsheets',
    'https://www.googleapis.com/auth/drive'
])
creds = flow.run_local_server(port=0)
open('token.json', 'w').write(creds.to_json())
"

This saves token.json — you won't need to do this again unless the token expires.

Connect to Claude Code

Add to your .mcp.json in the project root:

{
  "mcpServers": {
    "lead-gen-toolkit": {
      "command": "/absolute/path/to/.venv/bin/python",
      "args": ["/absolute/path/to/server.py"]
    }
  }
}

On Windows: ``json { "mcpServers": { "lead-gen-toolkit": { "command": "C:\\path\\to\\.venv\\Scripts\\python.exe", "args": ["C:\\path\\to\\server.py"] } } } ``

Restart Claude Code. Run /mcp to confirm lead-gen-toolkit is connected.

Usage

Once connected, just talk to Claude:

Scrape 30 roofing contractors in Denver CO and save to a sheet called "Denver Roofers"

Read the leads from [sheet URL] and push them to Instantly campaign [campaign_id]

Create an Instantly campaign called "Plumbers NYC | Free Estimate" and give me the campaign ID

Cost

| Component | Cost | |---|---| | Apify Google Maps | ~$0.01–0.02 per lead | | Instantly | your plan's sending limits | | Google Sheets | free |

100 leads ≈ $1.50–2.50 total.

File structure

server.py          # MCP server (the thing you're installing)
execution/
  gmaps_lead_pipeline.py   # Google Maps scraper + Sheet writer
.env               # Your API keys (never commit this)
credentials.json   # Google OAuth credentials (never commit this)
token.json         # Google OAuth token (never commit this)

Troubleshooting

"Google token is missing or expired" — re-run the auth command above.

"INSTANTLY_API_KEY not set" — add your key to .env. Get it from Instantly → Settings → Integrations → API Keys → select all:all scope.

Apify scrape returns 0 results — include the city in your query: "HVAC in Austin TX" not just "HVAC".