For AI agents and developers

India's restaurants, menus and photos, one call away

A free MCP server and REST API over the same data as this site. No key and no sign-up for reads. Every answer says where each fact came from, so your agent can be honest with its user.

MCP server

Add it to Claude, ChatGPT, Cursor or any MCP client as a remote server (Streamable HTTP):

https://indiarestros.com/mcp

Tools: search_restaurants, search_dishes, get_restaurant, get_menu, list_places, get_place, report_correction, about_dataset.

MCP details

REST API

JSON over HTTPS, OpenAPI 3.1 described, CORS open.

GET https://indiarestros.com/api/v1/restaurants?q=dosa&city=bengaluru
GET https://indiarestros.com/api/v1/dishes?q=paneer+tikka&city=pune

API reference openapi.json

The rules

Reads are free: 120 requests a minute per IP without a key.
Cite the restaurant URL and the source of the facts you use.
Users contact restaurants directly. Do not resell contact details in bulk.
Report what is wrong with report_correction. It makes the data better for everyone.

What you get back

Restaurants	Name, kind, cuisines, address, geo, phones, website, hours (+ open_now in IST), amenities, price band, cost for two, photos with alt text, menu summary, sources, confidence and completeness scores, permanent IRR- id.
Menus	Sections and items with rupee prices, veg marks, descriptions, canonical dish slugs; dish search across every menu.
Places	States, cities and areas with counts by kind and cuisine, centres and children.
Dataset	Counts, sources, licences and the ranking formula via about_dataset and /api/v1/stats.

Restaurants: reach the agents

Soon, restaurants can receive questions and table requests from AI agents directly by WhatsApp, email or webhook, for a small flat fee. Ranking stays free and neutral.

For restaurants

search_dishes({
  "q": "veg thali", "city": "pune",
  "max_price": 250, "veg": true
})