Zomato Restaurant Scraper - City, Cuisine, Contact MCP server

Zomato Restaurant Scraper

📦 Open source · MIT: github.com/AnshumanAtrey/zomato-restaurant-scraper

Scrape restaurant listings from Zomato by city and cuisine. Get structured records with name, phone, address, cuisines, rating, and cost for two — no Zomato account required.

Built for B2B sales teams, food-tech researchers, ghost-kitchen operators, lead-gen agencies, and competitor intelligence.

Quick start

1. Pick a City from the dropdown (Bangalore, Delhi NCR, Mumbai, Dubai, Singapore, etc. — 40 cities)
2. Pick a Listing Type (Best, All, Dinner, Breakfast)
3. Optionally pick a Cuisine filter (Biryani, Pizza, North Indian, Chinese, etc. — 24 options)
4. Set how many restaurants you want (default 50, up to 1000)
5. Click Run

Typical 50-restaurant scan completes in 3-7 minutes.

What you get

Each restaurant is pushed as one dataset record:

A final summary record reports restaurantsScraped, restaurantsFound, failures, duration, and sourceUrls.

Common use cases

Build a B2B restaurant lead list

Pick Mumbai, Best restaurants, maxRestaurants: 200. Export the CSV. You now have 200 verified phone numbers + addresses for outreach.

Scrape a specific cuisine in a specific city

Pick Hyderabad, All restaurants, Cuisine: Biryani, maxRestaurants: 100.

Scrape a city not in the dropdown

Use the Custom city field under "Advanced overrides". Type Tirunelveli, Erode, Coorg, Manali, Madurai, Varanasi — auto-converted to the right Zomato URL.

Scrape a category not in the dropdown

Use the Custom cuisine field. Momos, Kebab, Hyderabadi, Awadhi all work.

Scrape an exact Zomato URL you already have

Paste it into Custom URL under "Advanced overrides". Overrides every other dropdown.

Inputs

Main fields

• City — Pick from 40 Zomato cities across India, UAE, Qatar, Oman, Malaysia, Singapore, Philippines, Sri Lanka, Portugal
• Listing Type — Best restaurants (curated top picks), All restaurants, Dinner spots, or Breakfast spots
• Cuisine — Optional filter across 24 cuisine categories, or "Any" for no filter
• How many restaurants — Cap on results (1-1000). Start with 20-50 to verify, then scale up
• Scan depth:
• Quick — restaurant URLs + names only (fast, cheap, useful for bulk URL lists)
• Standard — full data: phone, address, rating, cost (default — recommended for B2B leads)
• Deep — full data with extra retries for blocked pages (slowest, highest yield)

Advanced overrides (optional)

• Custom city — Type any Zomato city not in the dropdown. Auto-slugified
• Custom cuisine — Type any Zomato cuisine category not in the dropdown
• Custom URL — Paste any Zomato listing URL to scrape it directly. Overrides everything above

Runtime

• Run timeout (seconds) — Hard kill after this many seconds (default 1800 = 30 min)

How it works

1. Listing page: navigates to the constructed Zomato URL, scrolls to load lazy content, extracts all restaurant detail URLs (handles both /info and /order URL patterns)
2. Detail pages: visits each restaurant page concurrently using fast HTTP fetches (Cheerio) — typically 5-10x faster than full browser rendering
3. Auto-fallback: if the primary URL returns 0 restaurants, automatically retries /{city}/restaurants as a fallback

Shortened. The full README is on GitHub.