Connect AI agents to Senserity

The Senserity MCP server lets AI assistants perform due diligence programmatically. Ask an agent to check a company, and it calls the Senserity API tools to fetch real risk intelligence.

What is the Model Context Protocol?

The Model Context Protocol (MCP) is an open standard that allows AI assistants to use external tools during a conversation. Rather than copying and pasting data between applications, an MCP-enabled agent can call Senserity tools directly — searching for companies, pulling risk summaries, generating reports, and monitoring alerts without leaving the conversation.

Available tools

senserity_search_companies

Search for UK companies by name or number. Returns matched companies with basic details.

senserity_get_company

Get the full profile of a company including registered address, SIC codes, and status.

senserity_get_risk_summary

Retrieve the overall risk score, letter grade, category breakdown, red flags, and trajectory.

senserity_list_insights

Get all individual insight test results for a company, grouped by risk category.

senserity_run_insights

Trigger a fresh risk assessment for a company. Consumes credits per existing rules.

senserity_get_alerts

Get risk alerts for monitored companies, filterable by severity and category.

senserity_add_to_watchlist

Add a company to the monitoring watchlist for continuous risk tracking.

senserity_generate_report

Generate a PDF due diligence report for a company.

senserity_list_reports

List existing reports, optionally filtered by company number.

senserity_toggle_monitoring

Enable or disable monitoring for a company on the watchlist.

Setup

Add the Senserity MCP server to your AI tool's configuration. The server supports two transport modes: stdio (local, via npx) and HTTP/SSE (remote, for tools that connect over the network).

Claude Desktop / Claude Code

Add to claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/, Windows: %APPDATA%\Claude\):

{
  "mcpServers": {
    "senserity": {
      "command": "npx",
      "args": ["senserity-mcp-server"],
      "env": {
        "SENSERITY_API_KEY": "sns_live_your_key_here"
      }
    }
  }
}

Cursor

Add to .cursor/mcp.json in your project root:

{
  "mcpServers": {
    "senserity": {
      "command": "npx",
      "args": ["senserity-mcp-server"],
      "env": {
        "SENSERITY_API_KEY": "sns_live_your_key_here"
      }
    }
  }
}

VS Code (GitHub Copilot)

Add to .vscode/mcp.json or your VS Code settings:

{
  "servers": {
    "senserity": {
      "type": "stdio",
      "command": "npx",
      "args": ["senserity-mcp-server"],
      "env": {
        "SENSERITY_API_KEY": "sns_live_your_key_here"
      }
    }
  }
}

Windsurf

Add to your Windsurf MCP configuration (Settings → MCP Servers → Add Server):

{
  "mcpServers": {
    "senserity": {
      "command": "npx",
      "args": ["senserity-mcp-server"],
      "env": {
        "SENSERITY_API_KEY": "sns_live_your_key_here"
      }
    }
  }
}

ChatGPT, Perplexity, and other remote clients

A remote HTTP/SSE endpoint for tools that require a network connection (ChatGPT Developer Mode, Perplexity, Cline, and custom agent frameworks) is on the roadmap. For now, the Senserity MCP server supports local stdio transport only.

Requirements

A Professional or Enterprise plan with API access enabled.
A valid API key with the appropriate scopes (at minimum companies:read and insights:read).
Node.js 18+ for the npx transport.

Example conversation

This is what a conversation looks like when an AI agent uses the Senserity MCP tools:

User

Can you check the risk profile for Tesco PLC?

Agent

Calling senserity_search_companies({search: "Tesco PLC"})...

Calling senserity_get_risk_summary({company_number: "00445790"})...

Tesco PLC (company number 00445790) has a risk grade of B with an overall score of 22.4 out of 100. The strongest areas are Governance (3.1) and Network (5.2). The weakest category is Cyber at 45.8, which is common for large organisations with complex web infrastructure. There are no active red flags.