Diliguard MCP Integration Guide

Diliguard natively supports the Model Context Protocol (MCP), allowing you to seamlessly integrate our corporate intelligence and OSINT modules directly into AI assistants like Claude Desktop, Cursor, and Windsurf.

Instead of writing scripts or manually pinging our API, you can now simply ask your AI:

"Run a complete vendor risk assessment on Google and summarize the findings in a table."

Don't have an account yet? Register for free here to get your API key.

How it Works

We provide an official Node.js MCP server package: @sluxia/diliguard-mcp. You can run this package locally using npx, and it will securely connect your AI assistant to your Diliguard account.

Because the package runs locally on your machine, your DILIGUARD_API_KEY is securely stored in your own environment variables and is never exposed in URLs or logged by external proxies.

Available AI Tools

Once connected, your AI assistant will automatically discover the following Diliguard tools:

1. run_corporate_kyc: Runs Domain, VAT, and IBAN checks to build a corporate risk profile.
2. run_person_trace: Runs Email, Phone, and Username checks for deep person profiling.
3. run_crypto_investigation: Runs Sanctions, Balance, and IP Intelligence checks for a given crypto wallet.
4. run_vendor_risk: Evaluates multiple companies concurrently to check for sanctions, UBOs, regulatory fines, and valid corporate registration.
5. run_ultimate_fraud: A holistic snapshot combining both Corporate and Person parameters (triggers all relevant modules).
6. run_aml_pep_screening: Checks global watchlists for Politically Exposed Persons (PEPs) and Anti-Money Laundering (AML) flags.
7. run_registry_check: Checks GLEIF and Ultimate Beneficial Ownership registries for deep corporate structures.
8. run_ip_intelligence: Investigates IP addresses for VPNs, Proxies, Datacenters, and network risks.
9. run_transaction_scoring: Evaluates transaction risk by combining IP Intelligence and Email Profiling into a unified fraud score.

Setup Instructions (Claude Desktop)

To connect your local Claude Desktop app to Diliguard's intelligence engine, update your claude_desktop_config.json file.

1. Open your Claude Desktop configuration:

   • Mac: ~/Library/Application Support/Claude/claude_desktop_config.json
   • Windows: %APPDATA%\Claude\claude_desktop_config.json

2. Add the Diliguard mcpServer entry using the npx command. Be sure to insert your own personal Admin Key:

{
  "mcpServers": {
    "diliguard": {
      "command": "npx",
      "args": ["-y", "@sluxia/diliguard-mcp"],
      "env": {
        "DILIGUARD_API_URL": "https://api.diliguard.com",
        "DILIGUARD_API_KEY": "YOUR_PERSONAL_ADMIN_KEY"
      }
    }
  }
}

3. Restart Claude Desktop. You should now see a small "plug" icon or "Diliguard" listed under your available MCP tools!

Testing with the MCP Inspector

If you want to debug the connection before adding it to Claude, you can use the official Model Context Protocol Inspector:

DILIGUARD_API_URL="https://api.diliguard.com" DILIGUARD_API_KEY="YOUR_KEY" npx @modelcontextprotocol/inspector npx -y @sluxia/diliguard-mcp

This will spin up a local web UI where you can view our tools and execute test payloads directly against the Diliguard backend.