# NYCfoodie

> The data layer for restaurant taste: structured editorial NYC restaurant recommendations for AI agents over MCP.

NYCfoodie is a free public MCP (Model Context Protocol) server. It exposes
5,767 New York City restaurant venues, 1,841 numeric ratings, 1,837 full
editorial reviews, 929 ranked guides, and 381 tags, sourced from professional
editorial coverage.

## Endpoint

- MCP (Streamable HTTP, no authentication): POST https://nycfoodie-production.up.railway.app/mcp
- Health check: GET https://nycfoodie-production.up.railway.app/healthz
- Source code and client setup: https://github.com/tireniajilore/nycfoodie
- Registered in the Official MCP Registry as io.github.tireniajilore/nycfoodie

## Tools

- search_restaurants: full-text search across venues, cuisines, neighbourhoods, occasions
- get_restaurant: full detail for one venue — review prose, numeric rating, booking intel
- compare_restaurants: side-by-side structured comparison of a shortlist
- find_guides: search ranked editorial guides (e.g. best Italian in the West Village)
- find_similar: venues similar to a given restaurant
- guide_consensus: how often a venue appears across guides (cross-guide consensus)
- top_rated: highest-rated venues by area or cuisine
- submit_feedback: rate a recommendation result

## Use cases

- "Find me a romantic Italian spot in the West Village under $70 a head"
- "Compare these three ramen places for a group dinner"
- "What do the guides say is the best pizza in Brooklyn right now"