Zomato MCP Server

What Zomato MCP Server MCP server does

Zomato MCP Server MCP server gives an AI client access to food-ordering workflows provided by Zomato. Its supported capabilities cover the path from discovering a restaurant to completing an order:

• Search for nearby restaurants using location and preferences.
• Browse menus with item prices, descriptions, and ratings.
• Create a cart, add items, and customize an order.
• Place food orders and receive order-tracking support.
• Complete payment through QR code integration.

Example requests include finding highly rated restaurants nearby, filtering for vegan or pizza options, adding a specified quantity of menu items, reordering from a previous order, and ordering a familiar meal.

How it works

The service is available at https://mcp-server.zomato.com/mcp. Claude can connect to that URL as a custom connector, while a manual Claude configuration can use npx mcp-remote to bridge the remote endpoint. Visual Studio Code can register the same address as an HTTP MCP server.

Authentication uses OAuth, and the project currently lists a limited set of approved redirect URIs. This restriction matters when selecting a client or configuring a custom OAuth flow.

Setup and configuration

For Claude, users with a subscription can add the hosted URL under Settings, Connectors, and custom connectors.

For Visual Studio Code, the documented options are a one-click installation link or a manual mcp.json entry. The manual entry uses the server name zomato-mcp-server, sets type to http, and points to the hosted endpoint. No environment variables or local service process are specified in the provided setup instructions.

Developer access is currently available by request for personal use. The README links to an access request form and says requests will be reviewed.

Tools and capabilities

Zomato MCP Server MCP server is suited to conversational food-ordering tasks that need location-aware discovery, menu details, cart changes, ordering, order tracking, and payment initiation through QR codes. The supplied examples show that an agent can interpret natural-language constraints such as distance, cuisine or dietary preference, item quantity, and previous-order requests.

The material describes service capabilities rather than a stable list of callable tool identifiers. Integrations should therefore avoid assuming specific function names unless the connected endpoint provides them during MCP discovery.

Limitations and notes

Zomato currently states that third-party applications cannot be built on top of Zomato MCP because of security and legal considerations. Integration discussions require contacting Zomato. The documented developer access is for personal use, and the project describes the integration as testing-only.

The OAuth redirect warning is an operational limitation: only the callback URLs listed in the README are currently whitelisted. The listed options include Claude, ChatGPT, Visual Studio Code, and Postman callback addresses. Clients outside those cases may require enablement before authentication works.

Zomato also disclaims liability for problems caused by erroneous behavior or an MCP integration that does not function. Developers should verify ordering and payment outcomes rather than treating an agent response as confirmation on its own.