Content
# @anujsup/milkbasket-mcp
[](https://www.npmjs.com/package/@anujsup/milkbasket-mcp)
[](https://www.npmjs.com/package/@anujsup/milkbasket-mcp)
[](https://opensource.org/licenses/MIT)
[](https://modelcontextprotocol.io)
[](https://nodejs.org)
> AI-powered grocery shopping for India — a [Model Context Protocol (MCP)](https://modelcontextprotocol.io) server for the [Milkbasket](https://www.milkbasket.com/) API.
Connect your AI assistant (Cursor, Claude Desktop, Windsurf, or any MCP client) to your Milkbasket account and let it handle your daily groceries — log in with OTP, browse products, build your cart, check your wallet, and review your full order summary, all through natural language.
**What you can do with this:**
- Ask your AI to find products — "Search for low-fat milk under ₹50"
- Build and manage your cart — "Add 2 Amul butter, remove the lassi"
- Check your order before delivery — "Show me my cart with the full bill"
- Keep track of your wallet — "What's my Milkbasket balance?"
- Fetch account details — "What city and hub am I assigned to?"
## Prerequisites
- Node.js 18 or higher
- A Milkbasket account (Indian mobile number)
- An MCP client (Cursor, Claude Desktop, etc.)
## Installation
### Option 1: npx (recommended, no install needed)
```json
{
"mcpServers": {
"milkbasket": {
"command": "npx",
"args": ["-y", "@anujsup/milkbasket-mcp@latest"]
}
}
}
```
### Option 2: Global install
```bash
npm install -g @anujsup/milkbasket-mcp
```
```json
{
"mcpServers": {
"milkbasket": {
"command": "milkbasket-mcp"
}
}
}
```
### Option 3: Local install
```bash
npm install @anujsup/milkbasket-mcp
```
```json
{
"mcpServers": {
"milkbasket": {
"command": "node",
"args": ["node_modules/@anujsup/milkbasket-mcp/dist/index.js"]
}
}
}
```
## Cursor setup
Add the config to `.cursor/mcp.json` in your project root (project-level) or `~/.cursor/mcp.json` (global):
```json
{
"mcpServers": {
"milkbasket": {
"command": "npx",
"args": ["-y", "@anujsup/milkbasket-mcp@latest"]
}
}
}
```
Reload Cursor after saving.
## Available Tools
### Authentication
| Tool | Description |
|---|---|
| `auth_request_otp` | Send OTP to your registered mobile number |
| `auth_verify_otp` | Verify OTP and log in |
| `auth_status` | Check if you are currently authenticated |
| `auth_logout` | Log out and clear stored tokens |
### User
| Tool | Description |
|---|---|
| `get_user_details` | Fetch your name, email, city, hub, and membership info |
| `get_wallet_balance` | Fetch your current Milkbasket wallet balance and cashback |
### Products
| Tool | Description |
|---|---|
| `list_products` | List featured/flash-deal products (paginated) |
| `search_products` | Search products by name or keyword (paginated) |
| `get_hero_collections` | Fetch hero banner collections shown on the home screen |
### Cart
| Tool | Description |
|---|---|
| `add_to_cart` | Add a new product or increase quantity in the cart |
| `remove_from_cart` | Decrease quantity or remove a product from the cart |
| `get_basket` | Lightweight cart snapshot — item count, product IDs, basic totals |
| `get_extended_basket` | Full cart view — product names, images, discounts, delivery fee, GST, final payable amount |
## Authentication Flow
```
1. auth_request_otp → { phone: "9XXXXXXXXX" }
2. Receive OTP via SMS
3. auth_verify_otp → { phone: "9XXXXXXXXX", otp: "XXXX" }
4. You are now logged in
```
## Security & Privacy
**This server runs 100% locally on your machine.** Your phone number, OTP, and tokens never leave your device — they are never sent to any third party, including the AI assistant. Here's exactly what happens:
- Your phone number is sent **directly** from your machine to Milkbasket's servers to request an OTP
- The OTP you share with the AI is passed **directly** from your machine to Milkbasket to verify login
- The AI model only sees the response (success/failure) — it cannot intercept or store your credentials
- Access tokens are stored in your **OS keychain** (macOS Keychain, Windows Credential Manager, Linux Secret Service) using `keytar` — the same secure storage used by apps like VS Code and 1Password
- If `keytar` is unavailable, an **AES-256-GCM encrypted file** is used as fallback
- You can log out at any time using `auth_logout`, which clears all stored tokens immediately
## Cart Flow
```
search_products → find productId and price
add_to_cart → { productId, price, quantity }
remove_from_cart → { productId, price, quantity: 0 } ← set 0 to delete
get_extended_basket → full checkout view
```
## What you can ask your AI
Once connected, just talk to your AI naturally:
| What you say | What happens |
|---|---|
| "Log me in to Milkbasket" | Sends OTP → verifies → stores token |
| "Search for paneer" | Calls `search_products` and returns matches |
| "Add 2 Amul Gold milk to my cart" | Calls `add_to_cart` with the right product |
| "Remove the lassi from my cart" | Calls `remove_from_cart` with quantity 0 |
| "Show me my cart with full bill" | Calls `get_extended_basket` with itemized bill |
| "What is my wallet balance?" | Calls `get_wallet_balance` |
| "What are my account details?" | Calls `get_user_details` — name, city, hub, membership |
| "What deals are on today?" | Calls `list_products` for flash deals |
| "Log me out" | Clears tokens from keychain |
## Example usage in Cursor
```
Search for Amul milk products
→ search_products({ query: "amul milk" })
Add 2 units to cart
→ add_to_cart({ productId: 1234, price: 28, quantity: 2 })
View full cart with bill breakdown (product names, discounts, delivery fee, GST, payable)
→ get_extended_basket()
Reduce quantity by 1
→ remove_from_cart({ productId: 1234, price: 28, quantity: 1 })
Remove product entirely
→ remove_from_cart({ productId: 1234, price: 28, quantity: 0 })
Check wallet balance
→ get_wallet_balance()
Get account info
→ get_user_details()
```
## License
MIT © [Anujsup](https://github.com/Anujsup)
MCP Config
Below is the configuration for this MCP Server. You can copy it directly to Cursor or other MCP clients.
mcp.json
Connection Info
You Might Also Like
AP2
AP2 provides code samples and demos for the Agent Payments Protocol.
google-meta-ads-ga4-mcp
MCP server for Google Ads, Meta Ads & GA4 — works with ChatGPT, Claude,...
nuwax
Nuwax AI enables easy building and deployment of private Agentic AI solutions.
Train-in-Silence
The first Task-Aware MCP server and automated VRAM calculator for LLM...
stacklit
108,000 lines of code. 4,000 tokens of index. One command makes any repo...
AppClaw
AI-powered mobile automation agent — describe what you want in plain...