@getvexi/mcp

npm version License: MIT Website

Structured business data for AI agents - connect Claude, Cursor, and any MCP-compatible client to the Vexi API in 30 seconds.

---

What is Vexi?

Vexi is a structured business data API designed for AI agents. Instead of scraping and parsing raw HTML, your agent gets a clean, typed Agent Business Object (ABO) - a JSON document with identity, offerings, contact channels, trust signals, and quality metadata.

With one call, agents can discover businesses, fetch full records, and request fresh crawls when data is missing or stale.

---

Prerequisites

• Node.js 18+
• A Vexi API key - get one free at https://getvexi.dev/signup

---

Installation

Claude Desktop

1. Open your Claude Desktop MCP configuration file:

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

2. Add or merge the vexi server entry under mcpServers.
3. Paste the following configuration (replace the API key with yours):

{
  "mcpServers": {
    "vexi": {
      "command": "npx",
      "args": ["-y", "@getvexi/mcp"],
      "env": {
        "VEXI_API_KEY": "vxi_live_..."
      }
    }
  }
}

Windows users: If npx doesn't work, use the cmd wrapper instead:

{
  "mcpServers": {
    "vexi": {
      "command": "cmd",
      "args": ["/C", "npx", "-y", "@getvexi/mcp"],
      "env": {
        "VEXI_API_KEY": "vxi_live_..."
      }
    }
  }
}

1. Save the file and fully quit Claude Desktop.
2. Relaunch Claude Desktop - the vexi tools should appear in your MCP tools list.

Cursor

1. Open Cursor.
2. Go to Settings > MCP.
3. Add a new server:

• Command: npx
• Args: -y, @getvexi/mcp
• Env: VEXI_API_KEY=vxi_live_...

4. Save and restart Cursor.

Other MCP clients

Run the server via stdio:

npx -y @getvexi/mcp

Set VEXI_API_KEY in the process environment before starting the client.

---

Available Tools

Tool	Description
search_businesses	Search for businesses by query and location
get_business	Get a full ABO for a specific business by slug
crawl_business	Queue a new business for crawling
list_categories	List all available business categories

---

Example Usage

User: Find the top 3 CRM platforms in the US.

Claude calls search_businesses:

{
  "query": "CRM platform",
  "location": "US",
  "limit": "3"
}

Claude: Here are three CRM platforms in the US:

Business	Completeness
HubSpot	0.94
Salesforce	0.97
Pipedrive	0.89

---

User: Give me full details on HubSpot.

Claude calls get_business:

{
  "slug": "hubspot"
}

Claude: Returns the full ABO for HubSpot - identity, offerings, contact channels, trust signals, operations, agent interface hints, and quality metadata in one structured document.

---

ABO Schema

Every business is returned as an Agent Business Object with eight top-level sections:

{
  "identity": {
    "id": "biz_hubspot",
    "name": "HubSpot",
    "slug": "hubspot"
  },
  "offerings": {
    "products": ["CRM", "Marketing Hub", "Sales Hub"],
    "price_range": "$$"
  },
  "location": {
    "headquarters": "Cambridge, MA",
    "country": "US"
  },
  "contact": {
    "website": "https://hubspot.com",
    "email": "support@hubspot.com"
  },
  "trust": {
    "rating": 4.5,
    "review_count": 12000
  },
  "operations": {
    "founded_year": 2006,
    "employee_range": "5000+"
  },
  "agent_interface": {
    "summary": "Inbound marketing and CRM platform for SMBs and enterprises",
    "suggested_queries": ["Compare pricing tiers", "List integrations"]
  },
  "quality": {
    "completeness_score": 0.94,
    "last_updated": "2026-05-28T12:00:00Z"
  }
}

---

Get your API Key

Sign up at https://getvexi.dev/signup.

Start free with 500 credits - no credit card required.

---

Links

• Website: https://getvexi.dev
• API Docs: https://getvexi.dev/docs
• Dashboard: https://getvexi.dev/dashboard
• npm: https://www.npmjs.com/package/@getvexi/mcp

---

License

MIT