Skip to content
MCP Server Directory logo
developer-tools

Main

Model Context Protocol (MCP) server for BatchData.io property and address APIs - Real estate data integration for Claude and other AI assistants

Reviewed overview

MCP server overview

What is the project about? BatchData - Real Estate - MCP Server is a Model Context Protocol (MCP) server that integrates BatchData.io property and address APIs, exposing address verification, geocoding, skip-tracing, and advanced property search/lookup functionality to Claude and other AI assistants or developer clients.

How to use the project?

  • Quick install for Claude via Smithery: npx -y @smithery/cli install @zellerhaus/batchdata-mcp-real-estate --client claude
  • Manual: git clone, npm install, create a .env with BATCHDATA_API_KEY, npm run build, npm start
  • Docker: use docker-manage.sh (test/run/logs) or docker build/run with an env file; docker-compose up -d supported
  • Configure Claude Desktop by adding an mcpServers entry pointing to the server script or docker exec command and set BATCHDATA_API_KEY in the env.

Key features

  • Address operations: verify-address (USPS standardization), autocomplete-address, geocode-address, reverse-geocode
  • Property operations: lookup-property (by address/APN), search-properties (rich filters), search-properties-by-boundary, count-properties
  • MCP-compliant server for integration with Claude and other MCP clients
  • Docker-first support: multi-stage builds, non-root user, health checks, resource limits, production-ready signal handling
  • TypeScript + runtime validation (zod), robust error handling for invalid keys, rate limits, malformed requests

Use cases

  1. Real estate agents/analysts: find comps, pull property details, run market counts or geographic searches
  2. Data cleansing: USPS address verification and standardization at scale
  3. Location-based applications: geocoding and reverse-geocoding for mapping and proximity searches
  4. Lead generation/skip tracing: validated addresses and owner lookups
  5. Embedding property data into AI assistants (Claude) for natural language queries and workflows

FAQ from the project

  • Q: Do I need an API key? A: Yes — set BATCHDATA_API_KEY (get from https://app.batchdata.com/settings/api).
  • Q: Is this compatible with Claude? A: Yes — the server is an MCP server and includes configuration examples for Claude Desktop (node or Docker-based).
  • Q: How do I run it in Docker? A: Use the included docker-manage.sh for build/run/test/logs or docker build -t batchdata-mcp-server:latest . and docker run --rm --env-file .env batchdata-mcp-server:latest. docker-compose is supported.
  • Q: What are API rate limits? A: Address verification up to 5,000 (1,000 recommended per batch), geocoding ~90 (75 recommended per batch), property search up to 1,000 requests; follow BatchData.io limits per endpoint.
  • Q: Is it production-ready and secure? A: It includes production-minded Docker features (multi-stage, non-root, health checks), error handling, and recommended resource limits, but you must secure the API key and host appropriately.
  • Q: What license? A: MIT.
  • Q: Common errors? A: Invalid/missing API key, exceeding BatchData rate limits, malformed requests, or network failures — the server returns descriptive errors for these conditions.