Skip to main content
Glama
restaurant-health-inspection-scores-mcp

restaurant-health-inspection-scores-mcp

restaurant-health-inspection-scores-mcp

Restaurant health inspection scores for AI assistants: the latest health inspection grade, score or result of a restaurant in New York City, Chicago, Austin or Boston, the violations the inspector cited, its inspection history, and the restaurants that recently failed or were closed in a neighborhood. Ask what grade Katz's Delicatessen has, whether a Chicago restaurant passed its last inspection, or which restaurants in ZIP 10002 were closed this month.

This is a Model Context Protocol (MCP) server. It reads each city's restaurant inspection dataset on the city open data portal at call time (New York City and Chicago update daily, Boston weekly, Austin with a lag of a few weeks). It needs no API key.

Tools

Tool

What it returns

restaurant_inspection_search

Find a restaurant by name (optionally city, ZIP or street) and get its latest inspection: NYC posted letter grade and score in points, Chicago Pass / Pass w/ Conditions / Fail, Austin score out of 100, Boston result and open violations. Searches all four cities when no city is given.

restaurant_inspection_history

One establishment by city and id (or name): the violations cited at the latest inspection, the NYC letter grade, past inspections, and how many inspections in the last 3 years failed or ended in a closure.

restaurant_inspection_failures

Recent failed inspections in a city, NYC borough or ZIP code: NYC closures and C-range scores, Chicago Fail results, Austin scores under 70, Boston failures and closures, over the last N days up to the newest published inspection.

restaurant_inspection_cities

Covered cities, how each one grades an inspection, what detail its data has, and the newest published inspection date in each dataset.

Every result has structured JSON, a plain-text summary, as_of, source and more (a page on mcpstamp.com). Every tool is marked read-only.

Related MCP server: nyc-restroom-mcp

What it knows that your model does not

  • Each covered restaurant's latest health inspection as published by the city this week: New York City letter grades and scores, Chicago pass/fail results, Austin scores and Boston results, not a training-time snapshot.

  • The violations cited at an inspection: NYC critical violations, Chicago inspector comments, Boston violations by level (*, **, ***) and whether they were corrected on re-inspection.

  • Which restaurants in a city, NYC borough or ZIP code failed an inspection or were closed by the Health Department in the last days or weeks.

Install

Hosted (custom connector, ChatGPT developer mode, any remote MCP client)

Paste this URL. No token, no sign-up.

https://restaurant-health-inspection-scores-mcp.lipmichal.workers.dev/mcp

Cursor

Add to ~/.cursor/mcp.json.

{
  "mcpServers": {
    "restaurant-inspections": { "url": "https://restaurant-health-inspection-scores-mcp.lipmichal.workers.dev/mcp" }
  }
}

VS Code

Add to .vscode/mcp.json.

{
  "servers": {
    "restaurant-inspections": { "type": "http", "url": "https://restaurant-health-inspection-scores-mcp.lipmichal.workers.dev/mcp" }
  }
}

Local stdio

The npm package is not published. Build it from this repository with Node 20 or newer, then point your MCP client at node /path/to/restaurant-health-inspection-scores-mcp/dist/stdio.js.

git clone https://github.com/theluckystrike/restaurant-health-inspection-scores-mcp && cd restaurant-health-inspection-scores-mcp
npm install && npm run build
node dist/stdio.js

Example prompts

  • "What health inspection grade does Katz's Delicatessen in New York have, and what violations did the inspector find last time?"

  • "Which restaurants in ZIP 10002 failed a health inspection or were closed in the last 30 days?"

  • "Did Au Cheval in Chicago pass its last health inspection?"

  • "Show me the inspection history of Neptune Oyster on Salem Street in Boston."

  • "What was Franklin Barbecue's last health inspection score in Austin?"

Data sources

Full list in DATA_SOURCES.md.

  • New York City: DOHMH New York City Restaurant Inspection Results, NYC Open Data (data.cityofnewyork.us, dataset 43nn-pn8j).

  • Chicago: Chicago Department of Public Health Food Inspections, Chicago Data Portal (data.cityofchicago.org, dataset 4ijn-s7e5).

  • Austin: Austin Public Health Food Establishment Inspection Scores, City of Austin Open Data (data.austintexas.gov, dataset ecmv-9xxi). It also covers the rest of Travis County that Austin Public Health inspects (Pflugerville, Lakeway, Bee Cave, Manor and others).

  • Boston: Inspectional Services Food Establishment Inspections, Analyze Boston (data.boston.gov, CKAN resource 4582bec6-2b4f-4f9e-bc55-cbaa73117f4c).

Limitations

  • Four cities only: New York City, Chicago, Austin and Boston. Montgomery County MD, Dallas and San Francisco were checked and left out because their open data stopped updating (2024, 2024 and 2019); King County WA requires a login.

  • Cities grade differently and the results are not comparable across cities: NYC scores are violation points (lower is better), Austin scores are out of 100 (higher is better), Chicago and Boston are pass or fail.

  • Austin's open data has scores only (no violation list), covers the last 3 years and is published a few weeks behind. NYC grades are posted only after certain inspections, so the latest score can differ from the posted grade.

  • Boston's result codes (HE_Fail, HE_FailExt, HE_VolClos, HE_TSOP, HE_Filed ...) are not documented by the city; the tools show the code next to a plain reading of it.

  • An inspection describes one day. A failure is often followed within days by a passing re-inspection.

  • The first read of a name in New York City takes 5 to 25 seconds, because the city portal scans about 295,000 rows; repeat calls are cached at the edge for an hour.

More

More MCP servers that answer questions your assistant cannot answer alone are on mcpstamp.com.

Part of the Zovo MCP collection at https://mcp.zovo.one

Built by theluckystrike.

Development

npm install
npm test                              # builds, then runs the fixture tests and an offline stdio test
node scripts/stdio-transcript.mjs     # live stdio round trip, writes .live/stdio-transcript.jsonl

CI (.github/workflows/bundle.yml) runs the tests, the type check and wrangler deploy --dry-run --outdir out, and uploads the Worker bundle as an artifact.

License

MIT

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • F
      license
      Not graded
      quality
      Not graded
      maintenance
      An AI-native restaurant discovery service that enables searching and receiving natural language recommendations for over 2,200 restaurants across 15+ US cities. It provides tools for accessing detailed restaurant info, curated lists, and cuisine-specific searches through the Model Context Protocol.
      -
    • A
      license
      Not graded
      quality
      C
      maintenance
      Enables AI agents to search restaurants, publish and maintain store data, submit verified freshness feedback, and query demand or updates through MCP tools. It provides both read-only and full administrative endpoints backed by open restaurant data.
      1
      AGPL 3.0
    • A
      license
      A
      quality
      C
      maintenance
      Enables AI clients to search and read over 2,400 New York City public datasets through the Socrata SODA API, with ready-made tools for restaurant health inspections and 311 service requests plus general dataset discovery and querying.
      4
      MIT