Claude MCP for lead generation: connect your AI client to a live prospect database

Most sales teams use Claude the same way: paste in a list of companies, ask it to write a personalized email. It works. But Claude is still flying blind — it knows nothing about your campaigns, your scored leads, or which prospects were disqualified last week.

Claude MCP for lead generation changes that. The same Claude session that drafts your outreach can now search for prospects across 11 platforms, check which campaigns are live, filter results by score, and export a custom Excel report — all without switching tabs.

What MCP is and why it changes prospecting

MCP — Model Context Protocol — is an open standard released by Anthropic in November 2024. It gives AI assistants a structured way to call external tools: databases, APIs, and services. The key difference from older integration methods is the permission model: you define exactly what the AI can read or write, and you can revoke access at any time without touching any code.

Before MCP, connecting an AI to an external system meant custom integrations, webhook chains, or general-purpose automation tools that weren't designed for conversational AI. With MCP, a sales rep adds Lode Leads as a connector in Claude and Claude immediately gains the ability to query, filter, and act on that live data — all through natural language.

For prospecting specifically, this is the difference between Claude knowing about your market and Claude actually working inside your pipeline.

Two ways to use Claude for sales — and why most teams stop at the first

Articles about Claude and lead generation typically describe one workflow: give Claude a list you exported from somewhere else and ask it to personalize outreach. Useful, but the lead list still has to come from somewhere. You're still doing the prospecting manually — Claude just handles the writing.

MCP adds a second workflow where Claude is the prospecting engine itself.

The workflow difference matters most for teams that already live in Claude. Adding Lode Leads as an MCP connector means prospecting becomes one more thing Claude handles — alongside search, CRM lookups, and writing — rather than a separate task in a separate tab.

How the Lode Leads MCP connection works

Setup is 3 steps:

1. Create a scoped API key in your Lode Leads dashboard. Pick a permission preset — Read-only Analyst, Operator, or Admin Automation — depending on what you want Claude to be able to do. The Operator preset is the recommended default for active prospecting.
2. Add the MCP server in Claude: Settings → Connectors → "Add custom connector". Name it anything you like and paste the URL: mcp.lodeleads.com/mcp?token=YOUR_TOKEN.
3. Confirm the connection by asking Claude: "What Lode Leads tools do you have access to?" Claude inventories every campaign, lead, and export tool your key permits — no manual configuration required.

From there, Claude can list active campaigns, check run status, filter leads by score or geography, launch new searches, and export reports. You control what it can touch through the permission preset you selected in step one.

For a concrete walkthrough of what this looks like end to end, see how Claude found 40 dermatology clinics in Cairo, qualified 26 of them against a 100-point scoring model, and built a 4-sheet scored Excel export — all through a single MCP session. That example also shows Claude automatically optimizing the campaign budget mid-session, cutting credit spend by 80% after the user asked it to reduce cost.

Full setup details and a step-by-step visual walkthrough are at lodeleads.com/mcp.

What you can ask Claude once connected

These prompts work once the MCP connection is live — no additional configuration, no code:

• "Find TikTok fitness creators in the UAE with 10k–100k followers who post regularly." Claude creates a TikTok campaign, runs it, and returns scored profiles with follower counts, engagement signals, and posting cadence.
• "Show me all leads from last week's LinkedIn campaign scoring above 80." Filtered live from your data — not a static export you pulled earlier.
• "Which of my active campaigns has the most qualified leads this week?" Cross-campaign comparison, answered from live run data.
• "Export the top 30 leads from my Google Maps campaign as Excel, include scores and disqualification reasons." Claude builds and downloads a 4-sheet file with lead data, score breakdowns, and a priority outreach tab.
• "Re-run my Cairo campaign but filter out any business with fewer than 10 reviews." Campaign modified and re-run with the new filter applied.

The significant change from dashboard-only use is context continuity. You move from "find leads" to "draft outreach for the top 3 scoring leads" to "schedule a follow-up search for next Monday" — all in the same Claude session, with Claude retaining the context of what you're working on throughout.

According to Salesforce's 2024 State of Sales report, sales reps spend only 28% of their time actually selling — the rest goes to administrative work, tool switching, and manual data tasks. Keeping prospecting and outreach drafting inside a single Claude session cuts the tool-switching overhead that eats into that 28%.

What each permission preset controls

When you generate an API key, you pick one of 3 presets. These aren't just labels — they define exactly which API actions Claude is allowed to call:

You can generate multiple API keys with different presets — one read-only key for a reporting assistant, another Operator key for the rep running campaigns. Revoking a key invalidates it immediately; there's no delay and no code change required. This matters for agencies handling multiple clients: each client engagement can have its own scoped key.

Who gets the most from this setup

Agencies managing multiple clients. One Lode Leads account holds separate campaigns per client. One Claude session with MCP access can query across all of them, compare campaign performance, and generate client-ready Excel exports — without switching platforms or dashboards.

Founders doing their own prospecting. No need to log into the Lode Leads dashboard between writing sessions. Ask Claude "any new qualified leads from my LinkedIn campaign this morning?" and get a filtered summary inline, then continue drafting in the same window.

SDR teams already running AI-first workflows. Reps who draft in Claude can add Lode Leads as one more connector — alongside CRM tools and search — and stay in a single interface throughout the prospecting-to-outreach cycle. Lead generation on 11 platforms with 100-point AI scoring becomes one more tool Claude has access to, rather than a separate system to log into.

How to get started

Frequently asked questions

Does Claude support MCP natively?

Yes. Claude's desktop app and claude.ai both support MCP connectors since the November 2024 release. You add them under Settings → Connectors — no code, no extensions required.

Can I use ChatGPT instead of Claude?

Any MCP-compatible AI client works with the Lode Leads MCP server. ChatGPT, Cursor, and other clients implementing the protocol are all valid options. Claude is the most direct setup today because Anthropic released and maintains the MCP standard.

How do I limit what Claude can do with my Lode Leads account?

Use the Read-only Analyst preset when generating your API key. Claude will be able to view and export data but cannot create or modify campaigns. You can also revoke any key from your dashboard immediately — no code change required.

Is this the same as the AI agent inside the Lode Leads dashboard?

The same 11-platform search engine and 100-point dual scoring model powers both. The difference is context: the in-dashboard agent keeps you inside Lode Leads. The MCP integration brings Lode Leads' tools into Claude — or any other MCP client — so you can prospect without leaving the AI environment you already work in.

Does MCP access cost extra?

MCP access is included on all plans, starting at $15/seat per month. Credits are only spent when campaigns run — browsing, filtering, and exporting data via MCP are free.