nycrentals

New York City homes, searched the way the city is actually organised: by borough and neighborhood. nycrentals is a thin, validating bridge to the hosted Borough MCP service — it exposes ten tools for finding rentals and homes for sale across Manhattan, Brooklyn, Queens, the Bronx and Staten Island, opening the listing or building behind a search result, and reading the market around a neighborhood before you commit to a viewing.

The bridge does two jobs. It serves the tool list locally, and it checks every call before anything leaves the machine: argument types, required fields, and the service's own house rules — pick one way of naming a place (areas, location or locations, never a mix), fee and tour flags that only ever switch on, no broker-fee flag on a sale search, at most ten area IDs in one search, and comparisons that stay between one and five neighborhoods. Calls that pass are forwarded over a single upstream session that is opened once and reused, and whatever the service answers — listings with prices, photos and freshness markers — comes back untouched.

The ten tools:

• search_rentals — active rentals across all five boroughs, filterable by price, beds and baths, square footage, amenities, building type, pet policy, transit score and internet speed, with no-fee, open-house and 3D-tour switches.
• search_sales — active for-sale listings with the same filters plus a sale-type selector.
• get_property — one listing in full: price, address, beds and baths, amenities, pet policy and building scores, in a compact form suited to chat clients.
• get_property_by_url — the same detail reached from a listing's URL path instead of its ID.
• get_building — a building's profile: type, year built, floors and units, amenities, policies, management and tax-abatement status.
• get_building_listings — the listings tied to one building, cheapest first.
• list_areas — the area directory itself: 316 neighborhoods and sub-areas, the five boroughs, and Northern New Jersey coverage, paginated.
• get_market_snapshot — current market statistics for a neighborhood.
• get_market_trends — recent price movement for a neighborhood, one row per date and event type.
• compare_neighborhoods — one to five neighborhoods side by side.

Area IDs follow the borough hundreds: Manhattan is 100, the Bronx 200, Brooklyn 300, Queens 400 and Staten Island 500, with neighborhoods numbered inside their borough.

Configuration

One credential, read from the environment:

export BOROUGH_API_KEY="..."

Every tool needs the key. Without it, calls fail immediately with BOROUGH_API_KEY is not set and nothing is sent anywhere. Access tiers are enforced by the hosted service, not by this bridge.

CLI

Each tool is a subcommand taking one JSON argument string:

npx @reneturcios/nycrentals search_rentals '{"location":"Brooklyn","maxPrice":3500,"noFee":true}'
npx @reneturcios/nycrentals get_property '{"id":"..."}'
npx @reneturcios/nycrentals compare_neighborhoods '{"locations":["Astoria","Jackson Heights"]}'
npx @reneturcios/nycrentals --version

Output is the forwarded result content as JSON. Exit codes: 0 on success, 1 on an error result (including a validation or upstream failure), 2 on a usage error.

MCP server

npx -y @reneturcios/nycrentals mcp

MCP client config:

{
  "mcpServers": {
    "nycrentals": {
      "command": "npx",
      "args": ["-y", "@reneturcios/nycrentals", "mcp"],
      "env": { "BOROUGH_API_KEY": "..." }
    }
  }
}

Docker

docker run -i -e BOROUGH_API_KEY="..." ghcr.io/cloudcorprecords/nycrentals
docker run --rm -e BOROUGH_API_KEY="..." ghcr.io/cloudcorprecords/nycrentals list_areas '{"level":1}'

Development

npm install
npm run build   # tsc -> dist/
npm test        # tsx --test test/*.test.ts (fake MCP upstream, no network)
npm run dev     # run the CLI from source

License

MIT — Copyright (c) 2026 Rene Turcios