Zomato’s new MCP (Mannequin Context Protocol) server makes it doable to order meals solely utilizing AI chat. The system was launched in 2025, and permits AI assistants like ChatGPT, Claude, and so on. to instantly work together with Zomato. In apply, it’s like merely say, “order a pizza close by,” and the AI does every part. On this information, we cowl what Zomato’s MCP server is and the way it works and stroll by way of setting it up step-by-step.
By the tip of this information, you’ll have a stable understanding of how Zomato’s MCP permits seamless “zero-UI” meals ordering with AI, and the best way to join it with different instruments like Claude Desktop or the Gemini CLI.
What’s the Zomato MCP Server?
Zomato’s MCP server is a Node.js implementation of the Mannequin Context Protocol, which exposes Zomato’s food-ordering APIs as conversational instruments. Principally, it is an interface between an AI assistant and Zomato’s providers. Zomato’s MCP server registers features, like searchRestaurants, getMenu, addToCart, and checkout in order that your AI can “name” them by way of chat.
For instance, for those who say:
“discover me pizza locations close to me,”
the backend server speaks to Zomato’s database and returns outcomes.
What’s Particular About Zomato’s MCP Implementation?
The Zomato MCP server exposes particular actions (instruments) for the AI to entry and communicates by way of both STDIO or HTTP/SSE. Within the case of Zomato’s MCP it will probably carry out actions like get_menu, add_to_cart, checkout and rather more. So when the AI makes a request (like making a request for the getRestaurantsForKeyword instrument), the exterior service server interprets the request into Zomato API calls and returns structured JSON information as a response.
This implies, for instance, that any AI mannequin suitable with the MCP commonplace will talk with Zomato’s system in your behalf.
Key Options of the Zomato MCP Server
The MCP server of Zomato permits the AI to leverage all meals ordering capabilities. Primarily, it offers your Assistant private ordering performance. The first features are as follows:
- Restaurant Discovery: Discover eating places in your space by location, delicacies sort, value tier, or score. For instance, you would possibly say “Discover North Indian Eating places below ₹400 close to me”, and in seconds see all of the choices.
- Menu Looking: Entry in-depth menus for any particular restaurant. The AI will have the ability to retrieve the merchandise names, description, costs, and choices for every merchandise. You would possibly say “Present me the menu for [Restaurant Name]”, and it is possible for you to to see the menu with out ever leaving the chat window.
- Cart Administration: Construct and edit your cart by way of voice. You need to use voice so as to add and take away gadgets and set up the small print (e.g., “Add Butter hen and a pair of rotis to my cart”). The MCP server will deal with the cart operations (together with modifiers and extras) behind the scenes.
- Order Checkout & Fee: As soon as the order is within the cart, the AI will have the ability to place the order and examine you out. It should apply any reductions and provides and create the fee request. The server will generate a UPI QR code so that you can pay at checkout.
Setting Up the Zomato MCP Server
To make use of the MCP server, you run an MCP shopper (native or distant) which is able to join your AI host to Zomato’s distant server. Beneath is a high-level setup course of:
- Node.js: You’ll need to have Node.js put in (model 16 or greater will work) so to run the MCP connector instrument. Use the command line to examine with `node -v`.
- Zomato Account: You’ll need a Zomato login to authorize the assistant. You don’t want something particular – simply your commonplace account login (it may be free or paid) that you just use when trying out.
- AI Host/Shopper: Decide the place the assistant goes to run. There are a lot of commonplace choices:
- Claude Desktop (which has native MCP help),
- ChatGPT (by way of the ChatGPT Connector or VS Code Copilot), or
- Gemini CLI (Google’s open-source terminal AI).
You can see a method so as to add the Zomato MCP server to the atmosphere you select.
Claude Desktop Configuration
Claude Desktop would be the best solution to strive the MCP server. As soon as Node.js is about up, you’ll use npx (a Node bundle runner) to entry, or run, Zomato’s MCP. Right here is the method:
- Edit Claude’s Config: Open the Claude Desktop utility and navigate to Settings > Developer > Edit config. This opens the claude_desktop_config.json file.
- Add Zomato Server: Within the JSON file, it’s essential add a server within the part known as “mcpServers”. For instance:
{
"mcpServers": {
"zomato-mcp": {
"command": "npx",
"args": [
"mcp-remote",
"https://mcp-server.zomato.com/mcp"
]
}
}
}- This tells Claude it must run npx mcp-remote https://mcp-server.zomato.com/mcp at any time when it accesses Zomato instruments.
- Save and Restart: When you save the file, you will want to give up Claude fully (Cmd+Q on a Mac or shut the applying from the taskbar on Home windows). When you reopen it, Claude will load Zomato MCP server.
- Authentication with Zomato: After finishing the above steps you might be able to run a command utilizing Zomato. When you’ve got by no means run a command (for instance, “discover pizza locations close to me”) utilizing Zomato, Claude gives you an authorization hyperlink. You’ll be able to click on the hyperlink, log in to your Zomato account, and end the method. Upon getting finished this, Claude will save the Oauth token that it acquired, and you’ll not must log again in once more. The GitHub readme states that the MCP server is utilizing Oauth for safe authentication.
Integrating with ChatGPT, Gemini CLI or VS Code
If you happen to use ChatGPT, terminal, or VSCode integrations, you may add the MCP server there as effectively:
Gemini CLI: Open ~/.gemini/settings.json Add a zomato-mcp entry below the mcpServers array:
"mcpServers": {
"zomato-mcp": {
"command": "npx",
"args": ["mcp-remote", "https://mcp-server.zomato.com/mcp"]
}
}Save the file and restart the Gemini CLI. Kind /mcp to substantiate the brand new server is listed. As earlier than, you’ll authenticate by way of the OAuth hyperlink.
VS Code: Go to to Zomato MCP repository. Scroll down and click on on Set up in VsCode and add the next mcp.json file or by clicking the “Set up in VS Code” button, level the url or command to https://mcp-server.zomato.com/mcp:
{
{
"servers": {
"zomato-mcp-server": {
"url": "https://mcp-server.zomato.com/mcp",
"sort": "http"
}
},
}Upon getting added this entry, you simply need to restart VS Code and authenticate by way of the hyperlink if you end up prompted. This is similar course of as with the Gemini CLI entry.
As soon as every part is configured accurately, you will notice zomato-mcp as a related server listed in your instrument, and you might be prepared to make use of it to execute instructions by way of your AI interface (e.g., utilizing /mcp in Gemini or typing your question in Claude, and so on.).
ChatGPT: Make the most of the official ChatGPT Connector. Within the ChatGPT settings for Connectors (Plugins), arrange a brand new MCP connector pointing to https://mcp-server.zomato.com/mcp. Confirm the redirect URL (chatgpt.com/connector_platform_oauth_redirect) is allowed.
Fast Verification
To substantiate the setup was profitable, try a easy check question after restart:
- In Claude/VSCode/Gemini, ask a easy immediate instance, “Present me vegetarian eating places close to my location.” The assistant ought to show a swift reply ready by Zomato.
- In Gemini CLI, you may sort the command /mcp checklist (or /mcp) to see the checklist of obtainable servers and examine to see if zomato-mcp seems.
If you’re authorization has been efficiently initialized, then the assistant won’t request a login once more for a similar account; in any other case, refelect to examine the redirect URLs and your JSON formatting.
Utilizing Zomato MCP: Workflow Examples
- Search Eating places: Ask a question much like “Discover North Indian eating places below ₹400 close to Gurugram”.
- Browse Menu: When you had chosen a restaurant, ask one thing like “Present me the menu for Themis Barbeque Home”.
- Add to Cart: To construct your order, immediate one thing like “Add 1 Butter Paneer Masala and three Butter Naan’s to my cart.”
Safety and Authentication
The safety of your interactions is managed utilizing OAuth. Whenever you first problem any Zomato immediate, the AI shopper will open a hyperlink that may allow you to securely log into your Zomato account and permit permissions. When you approve entry, the MCP server receives an entry token tied to your Zomato account. Claude or the Gemini CLI then shops the token so you may skip future logins. The server handles all authentication, and the AI by no means sees your Zomato password.
Funds are at all times safe. The MCP server by no means processes a bank card; it merely supplies a UPI QR code which you’ll scan along with your banking app.
Zomato has at the moment allowed particular redirect URLs (getting used within the Claude Desktop or ChatGPT plugin redirects, and so on.) to easily OAuth. In case you are utilizing a unique setup, you could want Zomato so as to add your particular redirect uri.
Conclusion
Zomato’s MCP server represents a revolutionary leap ahead for AI-powered meals ordering. With a fast command, clients can search, browse, cart and take a look at from Zomato’s ordering platform just by speaking with an AI assistant. As we’ve outlined, this know-how reveals how an app can fade into the background and permit AI to change into the consumer interface. It dramatically will increase the velocity of ordering and provides personalization and permits for seamless interoperability of AI techniques connecting each order.
The motion towards MCP servers like Zomato’s will eradicate the road between chatbots and actionable providers quickly. For Zomato, the MCP server units a precedent for taking a meals app and reworking it into an AI assistant.
Often Requested Questions
A. It exposes Zomato’s ordering APIs as MCP instruments. Your AI can search eating places, learn menus, add gadgets to cart, and take a look at by way of structured requests dealt with by the server.
A. You want Node.js, a Zomato account, and an AI shopper that helps MCP resembling Claude Desktop, ChatGPT with Connector, Gemini CLI, or VS Code.
A. Login occurs by way of OAuth. You approve entry in a browser, the AI shops an entry token, and your password isn’t shared. Funds use a UPI QR code for safe checkout.
Hey! I am Vipin, a passionate information science and machine studying fanatic with a powerful basis in information evaluation, machine studying algorithms, and programming. I’ve hands-on expertise in constructing fashions, managing messy information, and fixing real-world issues. My aim is to use data-driven insights to create sensible options that drive outcomes. I am desirous to contribute my abilities in a collaborative atmosphere whereas persevering with to study and develop within the fields of Knowledge Science, Machine Studying, and NLP.
Login to proceed studying and luxuriate in expert-curated content material.
