Vendr MCP Server

The Vendr MCP server gives AI assistants access to Vendr's proprietary software pricing catalog, which is derived from thousands of real SaaS purchase quotes. Procurement, finance, and IT teams can use it to look up software vendors and products, request fair-price estimates for a defined scope of work, and pull negotiation guidance backed by Vendr's benchmark data.

The server is published as an official open-source repository under the vendrinc GitHub organization and is also offered as a hosted endpoint at https://mcp.vendr.com/mcp. Vendr provides three integration paths: a one-click Claude Desktop extension (.dxt), a self-hosted local install from the GitHub repo for clients like Cursor or ChatGPT, and direct HTTP configuration against the hosted MCP endpoint with an X-API-Key header.

It exposes three tools that map directly to Vendr's public Catalog, Scope, and Pricing APIs. Access requires a Vendr API key, which is gated and requested by emailing Vendr. The Vendr API enforces a rate limit of 250 requests per minute and a quota of 150,000 requests per day.

Tools

Prerequisites

• A Vendr API key. Request one by emailing Vendr (see developers.vendr.com). The same key is used for both the REST API and the MCP server.
• Vendr API limits: 250 requests/minute and 150,000 requests/day per key.

Option 1: Claude Desktop extension (.dxt)

Download the prebuilt extension from the GitHub Releases page and drag it into Claude Desktop's Extensions window:

https://github.com/vendrinc/vendr-mcp/releases

You will be prompted to paste in your Vendr API key during install.

Option 2: Remote MCP endpoint (Claude, Cursor, custom clients)

Point your MCP client at Vendr's hosted endpoint with an X-API-Key header:

{
  "mcpServers": {
    "vendr": {
      "url": "https://mcp.vendr.com/mcp",
      "headers": {
        "X-API-Key": "YOUR_VENDR_API_KEY"
      }
    }
  }
}

Option 3: Local install from source

git clone https://github.com/vendrinc/vendr-mcp.git
cd vendr-mcp
npm install
npm run build
npm run start

The server is built with @modelcontextprotocol/sdk and is configured via a Context object:

import { makeServer } from "./src/index";

const context = {
  apiKey: "your-api-key",
  baseUrl: "https://api.vendr.com",
  userIdentifyingHeaders: {
    "x-vendr-end-user-email": "user@example.com"
  }
};

const server = makeServer(context);

The optional userIdentifyingHeaders enable per-end-user personalization when embedding the server in a multi-user application.

• Pull a fair-price benchmark for a renewal before walking into a vendor negotiation, using getCustomPriceEstimate with the exact scope (seats, tier, term length).
• Generate a one-page negotiation brief for an upcoming SaaS purchase by combining searchCompaniesAndProducts results with getNegotiationInsights.
• Triage inbound procurement requests in Slack: an agent looks up the vendor in Vendr's catalog and returns expected price ranges before a buyer is assigned.
• Build internal procurement copilots that let employees self-serve "is this quote reasonable?" checks against Vendr's benchmark data.
• Power finance team workflows that flag renewals where the proposed price is materially above Vendr's estimated fair price.

• "Search Vendr for Datadog and list the available product families and add-ons."
• "Give me a custom price estimate for 250 Salesforce Sales Cloud Enterprise seats on a 2-year term."
• "What negotiation levers should I use when renewing Snowflake? Pull insights from Vendr."
• "I have a quote for Notion at $18/user/month for 400 users. Compare it to Vendr's benchmark and tell me if it's high."
• "Find all observability vendors in Vendr's catalog and rank them by typical contract size."

• Official server from Vendr (the vendrinc GitHub org), so the tools track Vendr's underlying Catalog, Scope, and Pricing APIs.
• Offered three ways: a one-click .dxt extension, a hosted remote MCP endpoint, and a self-hostable open-source repo.
• Backed by Vendr's proprietary benchmark dataset built from real SaaS purchase quotes, which is hard to replicate elsewhere.
• Supports per-end-user identifying headers, making it usable inside multi-tenant agent applications.

• API access is gated: you must email Vendr to get a key, there is no self-serve signup.
• Only three tools are exposed. Lower-level Catalog, Scope, and Webhooks API surface area is not all wrapped in MCP tools.
• Rate limited to 250 requests/minute and 150k/day per key, and attribution requirements apply when surfacing Vendr data to end users.

• Azure MCP Server pricing tools for cloud (not SaaS) cost lookups.
• Building directly against the Vendr REST API without MCP, for custom integrations.
• Internal scripts against G2 or Gartner pricing data, though these lack Vendr's quote-derived benchmarks.