Swiggy MCP Server

Give your AI assistant live access to Swiggy (India's largest food‑delivery platform). This actor runs a Model Context Protocol (MCP) server so agents like Claude, ChatGPT, Cursor, and n8n can search restaurants, fetch full menus with prices, search dishes, and geocode locations on demand — no scraping code required.

It also runs as a classic scraper (input → dataset) when you just want a batch of rows.

Who it's for

• AI‑assistant & agent builders who need live Indian food‑delivery data in their tools.
• Analysts doing ad‑hoc restaurant/menu lookups from Claude or Cursor.
• SaaS teams embedding Swiggy data into LLM workflows.

MCP tools

Tool	Description
search_restaurants	Search or list restaurants near an Indian location — rating, cuisines, cost for two, delivery ETA, offers.
get_menu	Fetch a restaurant's full menu (categories → items with prices, veg flags, availability) by ID or URL.
search_dishes	Find restaurants matching a dish keyword near a location.
geocode	Resolve an Indian place/address to coordinates.

Connect your AI assistant

1. Start the actor in Standby mode (default) — it serves MCP over Streamable HTTP.
2. Point your MCP client at the actor's standby URL + /mcp:

{
    "mcpServers": {
        "swiggy": {
            "url": "https://<your-actor-standby-url>/mcp"
        }
    }
}

3. Ask your assistant things like "Find top‑rated biryani places in Koramangala and show the menu for the first one."

Example tool call

// search_restaurants
{ "location": "Bangalore", "query": "pizza", "limit": 10 }

// get_menu
{ "restaurant": "12808", "location": "Bangalore" }

Classic (dataset) mode

Run it normally with an input to push restaurant rows to a dataset:

{ "location": "Bangalore", "searchQuery": "biryani", "maxRestaurants": 20 }

Use cases

• Conversational food discovery inside an AI product.
• Agentic price/menu lookups for a specific restaurant on demand.
• Location‑aware restaurant search as a tool in a larger agent workflow.

FAQ

Which countries? India only — Swiggy's coverage.

Do I need an API key? No. The actor handles Swiggy's anti‑bot challenge automatically.

What transport does the MCP server use? Streamable HTTP at /mcp.

Is the data live? Yes — every tool call queries Swiggy in real time.

Is this legal? It accesses publicly available data; use the output in line with Swiggy's terms and applicable laws.