MCP Server: LeadMagic in Your AI Assistant

An AI model by itself can think and reason, but it cannot actually look up a prospect's email, verify whether an address is deliverable, or tell you what tech stack a company runs. It can only work with what is already in its context window — which means stale training data, not live business intelligence.

The LeadMagic MCP Server changes that. MCP (Model Context Protocol) is an open standard that lets AI assistants call external tools mid-conversation. Our server at mcp.leadmagic.io gives your AI tool 16 enrichment tools it can call on demand — validate an email, find a contact, enrich a company, search for job postings — without you leaving the chat.

What Is MCP and Why Does It Matter?

If you have used Zapier or Make, the concept is familiar. Those tools connect apps so data flows between them automatically. MCP does something similar but instead of you building the automation, your AI assistant decides what tools to use and when.

Here is the practical difference. Without MCP, if you ask Claude "find the VP Sales at Stripe," it will give you a best guess from its training data — which might be months old and could be wrong. With LeadMagic's MCP server connected, Claude calls the find_role tool, queries our live database, and returns the current name, verified email, phone number, and profile URL. Real data, not a guess.

The same applies in Cursor, VS Code, ChatGPT, Windsurf, and every other MCP-compatible client. Your AI assistant becomes a B2B data powerhouse that can act on live information.

16 Tools, One Server

Every tool runs through a single endpoint at https://mcp.leadmagic.io/mcp. Your AI assistant picks the right one automatically based on what you ask.

Finding people:

• find_email — Name + domain to verified work email
• find_mobile — B2B profile or email to direct mobile number
• find_role — Find the VP Sales, CTO, etc. at any company
• social_to_email — B2B profile URL to verified work email
• find_employees — Company to names, titles, emails, profiles

Validating data:

• validate_email — SMTP + MX check: is this address real?
• detect_job_change — Check if a contact switched companies

Enriching companies:

• search_company — Domain to industry, size, revenue, HQ, and 40+ fields
• search_profile — Full profile with title, company, history, education
• get_technographics — Domain to tech stack (CRM, analytics, hosting)
• find_competitors — Company to list of competitor companies
• get_company_funding — Funding rounds, investors, amounts raised

Searching:

• find_jobs — Search job postings by company, title, location

Account management:

• check_credits — Check your remaining credit balance (always free)

30-Second Setup

The entire setup is one JSON config block pasted into your AI tool's settings. Here is the universal config:

{
  "mcpServers": {
    "leadmagic": {
      "url": "https://mcp.leadmagic.io/mcp",
      "headers": {
        "x-leadmagic-key": "YOUR_API_KEY"
      }
    }
  }
}

Replace YOUR_API_KEY with your key from app.leadmagic.io/settings/api, save the config file, and restart your AI tool. That is it.

The config file location varies by tool:

• Cursor — .cursor/mcp.json (project) or ~/.cursor/mcp.json (global)
• Claude Desktop — ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\ (Windows)
• VS Code / Copilot — .vscode/mcp.json (workspace) or User profile
• Windsurf — ~/.codeium/windsurf/mcp_config.json
• OpenCode — opencode.json (project) or ~/.config/opencode/config.json (global)

We provide tested configs for 13 AI tools total, and any MCP-compatible client works with the same format.

For step-by-step instructions for your specific tool, see our integration guides: Cursor, Claude, ChatGPT, or OpenCode.

What You Can Ask

Natural language in, structured data out. Your AI picks the right tool automatically.

"Find the VP of Engineering at Stripe" — AI calls find_role, returns name, verified email, profile, and phone number.

"Validate these 50 emails and tell me which ones bounced" — AI calls validate_email for each address, returns a clean/bounced breakdown.

"What tech stack does notion.so use?" — AI calls get_technographics, returns CRM, analytics, hosting, marketing tools, and more.

"Enrich acme.com — I need everything" — AI calls search_company, returns 40+ fields: revenue, headcount, funding, industry, social links, and more.

"Has Jane Smith changed jobs since we last spoke?" — AI calls detect_job_change, returns current company and role with change date if applicable.

Who This Is For

MCP is not just for developers. Anyone who uses an AI tool daily can benefit from live B2B data inside their workflow.

Developers and Growth Engineers

You are building inside Cursor or VS Code. Mid-conversation, you need to verify an email format, look up a company's API capabilities, or check if a lead's data is still current. Instead of switching to a browser and manually querying, just ask your AI assistant. The data comes back inline, and you keep coding.

Sales Teams Using Claude Desktop

You are a sales manager or AE who uses Claude Desktop as a daily productivity tool. Before a meeting, ask Claude to pull the prospect's company details, recent funding, tech stack, and key contacts. Claude calls LeadMagic's tools and builds you a research brief in 30 seconds — no CRM tab-switching, no manual lookups.

Agencies Running Prospecting at Scale

You manage outbound campaigns for multiple clients using AI tools. Connect LeadMagic's MCP server once, and every campaign benefits from live enrichment. Ask your AI to validate a batch of emails before importing to Smartlead, find mobile numbers for a target account list, or identify which companies in a prospect list recently raised funding.

Vibe Coders and AI-Native Builders

You use Cursor or Claude to build products and workflows at speed. Having live B2B data available mid-conversation means your AI assistant can validate assumptions, look up real companies, and test with actual data instead of placeholders. Build a lead scoring model? Ask your AI to pull real company data to test it against.

Same Key, Same Credits

The MCP server uses the same API key and credits as the CLI and REST API. No separate account, no additional subscription. check_credits is always free — ask your AI how many credits you have left at any time.

If you prefer the command line for bulk processing, the LeadMagic CLI gives you CSV enrichment, AI-powered data querying, and full automation from your terminal. Same key, same credits, different interface.

Get Started

1. Get your API key (free credits included)
2. Paste the JSON config into your AI tool's MCP settings
3. Restart your AI tool
4. Ask it to "find the VP Sales at [any company]"

From config to first result in under a minute. Your AI assistant just became a B2B data powerhouse.

Check out the MCP Server product page for the full setup widget with tool-specific configs, or read our integration guides for Cursor, Claude, ChatGPT, and OpenCode.