Mirror Real Estate MCP Server
Connect Claude, ChatGPT and any other AI assistant directly to Mirror Real Estate's live property data through the open Model Context Protocol. Public, read-only, and free to use.
What you can do with it
The server exposes five tools backed by the same data that powers this website. An AI assistant connected to it can answer questions like "Find houses for sale in Toronto under $1M with 3+ bedrooms" or "Which cities have the most listings right now?" using real, current data.
search_listings
Search public listings by city, country, price, beds/baths, rent vs. sale, and more.
get_listing
Pull the full public detail for a single listing by its slug.
count_listings_by_city
List cities, grouped by country, with their public listing counts.
search_agents
Search public agents and brokers by keywords, city, or property type.
get_agent
Pull the full public detail for a single agent or broker by slug.
Connect in under a minute
Claude Code (CLI)
claude mcp add --transport http mirror https://www.mirrorrealestate.com/mcpThen run claude mcp list to confirm the five tools are available.
Claude Desktop
Add this to your claude_desktop_config.json
(Settings → Developer → Edit Config), then restart the app:
{
"mcpServers": {
"mirror": {
"type": "http",
"url": "https://www.mirrorrealestate.com/mcp"
}
}
}Any other MCP client
Point your client at https://www.mirrorrealestate.com/mcp using the Streamable HTTP
transport. No headers or credentials are required.
What's exposed (and what isn't)
Every tool is strictly read-only — nothing can be created, updated, or deleted. Results are mapped onto a slim allow-list of fields, so only data that is already public on this website ever leaves the server.
Internal database IDs, source/MLS numbers, and like/view counts are deliberately never returned. The endpoint is rate-limited to 60 requests per minute per IP.