How to Connect Claude Desktop to a Live B2B Lead Enrichment MCP API

Description and use

Connecting a MCP server to Claude Desktop enables agents to fetch verified B2B data in real time, eliminating hallucinations and saving costs with pay‑per‑success billing.

Problem

AI agents in local LLM environments often hallucinate or lack real-time firmographic and intent data, limiting effective B2B outreach.

Outcome

Integrating the MCP server eliminates hallucinations, provides verified B2B data, and introduces risk‑free metered billing, enabling cost‑effective scaling of AI‑driven SDRs.

Workflow

• 1. Obtain a free API key from https://lead-enrichment-mcp.agent-infra.workers.dev
• 2. Add the MCP server configuration to claude_desktop_config.json or Cursor’s MCP settings, specifying the command, args, and environment variables (ENRICHMENT_API_KEY and MCP_ENDPOINT).
• 3. In the LLM prompt, invoke the enrich_lead tool with a valid domain or email.
• 4. Receive a structured JSON‑RPC response containing companyName, industry, headcount, technographics, intentSignals, confidenceScore, and status.
• 5. Use the returned data to generate outreach messages or score leads.
• 6. Monitor billing: only successful enrichments with confidenceScore > 0.6 incur cost; failures cost $0.
• 7. Iterate and scale the agent swarm as needed.

Evidence

The server implements strict Zod-annotated schemas for every tool, ensuring that the model strictly adheres to expected parameter types

Source

Your account is only debited for successful enrichments that return a Confidence Score > 0.6. If the engine fails to locate the lead or returns low-confidence data, the query cost is exactly $0

Source