MCP Server

An MCP server enabling AI agents to extract structured web data using Nimble Technologies.

Nimble Labs Feature

The features included in the "Labs" section are experimental and provided for evaluation purposes only. These features may not be fully tested, stable, or integrated with all existing functionality. Future versions may differ from the current implementation.

If you are interested in using any of these experimental features in a production environment, please reach out to our team for guidance and support. Your feedback is valuable in helping us improve these features, and we welcome any issues, suggestions, or feedback you may have.

Most AI agents are blind. They hallucinate, they guess, they fail. All because they can't see the real world.

Model Control Protocol (MCP) is an open standard developed by Anthropic. It gives AI agents a simple, standardized way to plug into tools, data, and services.

The Nimble MCP Server bridges this gap by providing real-time, structured web data that transforms how agents interact with the digital world. With tools for web search, location data, and business intelligence, it turns chaotic web information into signals your agents can actually use.

Available Tools

• nimble_deep_web_search: Perform web searches across multiple search engines (Google, Bing, Yandex) with configurable parameters

• nimble_google_maps_search: Search for places on Google Maps and retrieve structured location data

• nimble_google_maps_reviews: Collect reviews for a specific place from Google Maps

• nimble_google_maps_place: Retrieve detailed, dynamic information about a specific place from Google Maps, including core details, metadata, amenities, and more.

Quick Start

The fastest way to use this MCP server is to connect to our official remote instance using Server-Sent Events (SSE) transport. This allows you to use the server without installing it locally.

Remote Connection Setup (SSE)

1. Get a Nimble API Key:

   • Register for an account at Nimble's signup page

   • Navigate to the "Pipelines" page and access the "Nimble API" pipeline

   • Obtain your API credentials (provided as a base64 token)

2. Prerequisites:

   • Ensure you have Node.js and npm installed on your system

   • The configuration below uses npx to run the MCP remote client without needing a local installation

3. Add Configuration to your MCP-Compatible Client:

   This configuration works with Claude Desktop, Cursor, Qodo, and any other client that supports Model Context Protocol (MCP).

   Add this configuration to your claude_desktop_config.json:

   {
     "mcpServers": {
       "nimble-mcp-server": {
         "command": "npx",
         "args": [
           "-y", "mcp-remote@latest", "https://mcp.nimbleway.com/sse", 
           "--header", "Authorization:${NIMBLE_API_KEY}"
           ],
         "env": {
           "NIMBLE_API_KEY": "Basic XXX"
         }
       }
     }
   }

   Note: Replace Basic XXX with your actual Nimble API key.

4. Start using the MCP tools with your AI agent!

Example using Nimble Maps API in Claude Desktop

We can't wait to see what cool things your AI agents will do with our tools! Have ideas for making these tools better or need help getting started? Drop us a line - we're here to help your agents get the data they need!