parknav_mcp_server

Bejarano03/parknav_mcp_server

3.1

If you are the rightful owner of parknav_mcp_server and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.

ParkNav MCP Server is a specialized server designed to facilitate the integration and management of parking navigation services using the Model Context Protocol (MCP).

Parknav AI

Parknav is a multi-agent system designed to find and enrich parking information. It uses a specialized agent to collect basic parking data and another to perform targeted web searches, providing detailed information like hours, pricing, and operator details to a frontend application.

Features

  • Intelligent Data Collection: Uses the Overpass API to find parking spots within a specified radius.
  • Automated Enrichment: An agent automatically performs web searches to find crucial details for each parking location.
  • Structured Data Output: All enriched data is stored in a PostgreSQL database and can be retrieved as a GeoJSON object.
  • Multi-Agent Architecture: The system uses two specialized agents

Getting Started

Prerequisites

  • Node.js (v8 or higher)
  • PostgreSQL Database/Neon
  • OpenAI API Key
  • SerpAPI API Key

Installation

  1. Clone Repo

  2. Installed dependencies via

    npm install

  3. Set up environment variables by creating a .env file in the root directory and adding the following:

    NEON_DB_URL="<YOUR_POSTGRESQL_CONNECTION_STRING>"
OPENAI_API_KEY="<YOUR_OPENAI_API_KEY>"
SERPAPI_API_KEY="<YOUR_SERPAPI_API_KEY>"

Running server

  1. Build mcp server via

    npm run build

  2. Run via

    node build/index.js