shop-mcp
MCP server for automating personal Amazon.in and Flipkart accounts via browser sessions, enabling product search, cart management, and guarded checkout.
README
<p align="center"> <img src="https://img.shields.io/badge/MCP-Server-blue?style=for-the-badge" alt="MCP Server"/> <img src="https://img.shields.io/badge/Python-3.10+-green?style=for-the-badge&logo=python" alt="Python"/> <img src="https://img.shields.io/badge/Playwright-Browser-orange?style=for-the-badge&logo=playwright" alt="Playwright"/> <img src="https://img.shields.io/badge/Sites-Amazon.in%20%7C%20Flipkart-purple?style=for-the-badge" alt="Sites"/> </p>
<h1 align="center">shop-mcp</h1>
<p align="center"> <b>A local MCP server for personal shopping automation on Amazon.in & Flipkart</b><br/> Search products, compare prices, manage cart, and place orders — all through your AI assistant with safety guards built in. </p>
<p align="center"> <a href="#-quick-start">Quick Start</a> • <a href="#-architecture">Architecture</a> • <a href="#-tools">Tools</a> • <a href="#-safety-guards">Safety Guards</a> • <a href="#-mcp-client-config">Client Config</a> </p>
How It Works
graph LR
A[AI Assistant<br/>Claude / Open AI / Kiro / Any AI Chat or LLM Model] -->|MCP Protocol<br/>stdio| B[shop-mcp<br/>Server]
B -->|Playwright<br/>Browser Automation| C[Amazon.in]
B -->|Playwright<br/>Browser Automation| D[Flipkart]
B --- E[Saved Sessions<br/>cookies only<br/>no passwords]
style A fill:#4A90D9,color:#fff
style B fill:#2ECC71,color:#fff
style C fill:#FF9900,color:#fff
style D fill:#F7D02C,color:#000
style E fill:#95A5A6,color:#fff
flowchart LR
AI["🤖 YOUR AI ASSISTANT\nClaude Desktop · Kiro · Any MCP Client"]
subgraph SERVER["🖥️ shop-mcp SERVER"]
direction TB
RO["📖 Read-Only\nsearch_product\nview_cart\nget_product_details\nget_order_status"]
MU["✏️ Mutations\nadd_to_cart"]
GC["🔐 Guarded Checkout\ncheckout_dry_run\nplace_order\nToken + ₹3000 cap\n5 min expiry"]
end
subgraph SITES["🌐 Shopping Platforms"]
direction TB
AMZ["🟠 Amazon.in\nSession cookies\nManual login once"]
FK["🔵 Flipkart\nSession cookies\nManual login once"]
end
AI -->|"MCP Protocol\nstdio"| SERVER
SERVER -->|"Playwright\nChromium"| SITES
style AI fill:#1a73e8,stroke:#0d47a1,stroke-width:3px,color:#fff
style SERVER fill:#e8f5e9,stroke:#2e7d32,stroke-width:2px,color:#2e7d32
style RO fill:#c8e6c9,stroke:#2e7d32,stroke-width:2px,color:#1b5e20
style MU fill:#fff9c4,stroke:#f9a825,stroke-width:2px,color:#e65100
style GC fill:#ffcdd2,stroke:#c62828,stroke-width:2px,color:#b71c1c
style SITES fill:#f3e5f5,stroke:#6a1b9a,stroke-width:2px,color:#6a1b9a
style AMZ fill:#fff3e0,stroke:#e65100,stroke-width:2px,color:#bf360c
style FK fill:#e8eaf6,stroke:#283593,stroke-width:2px,color:#1a237e
Key Features
| Feature | Description |
|---|---|
| Price Comparison | Search the same product on both platforms and compare prices instantly |
| Cart Management | View, add to cart across Amazon.in and Flipkart |
| Guarded Checkout | Orders require a dry-run summary + confirmation token + price cap |
| No Password Storage | You log in manually once; only session cookies are saved |
| Headed Browser | Visible Chromium window — no stealth, no detection issues |
| MCP Standard | Works with any MCP-compatible AI client via stdio transport |
Quick Start
# 1. Clone & setup
git clone https://github.com/detushar/shop-mcp.git
cd shop-mcp
python -m venv .venv
.venv\Scripts\activate # or source .venv/bin/activate on Linux/macOS
# 2. Install dependencies
pip install -r requirements.txt
playwright install chromium # downloads browser binary (~180MB)
# 3. Login to sites (one-time, manual)
python -m browser.login amazon
python -m browser.login flipkart
# 4. Run the MCP server
python server.py
Note:
playwright install chromiumdownloads the Chromium binary that Playwright controls. This is separate from the Python package and required for the server to work.
Tools
| Tool | Effect | Risk Level |
|---|---|---|
search_product |
Search a site for products | Read-only |
get_product_details |
Get price, title, availability for a product URL | Read-only |
view_cart |
List current cart contents | Read-only |
get_order_status |
Look up an existing order by ID | Read-only |
add_to_cart |
Add a product to the cart | Reversible |
checkout_dry_run |
Load checkout summary WITHOUT submitting | Read-only |
place_order |
Submit the order | Irreversible — guarded |
Safety Guards
place_order() will REFUSE if:
✗ No checkout_dry_run() was called first
✗ confirm_token doesn't match the latest dry-run
✗ Token is older than 5 minutes (stale prices)
✗ Order total exceeds ₹3,000 (configurable cap)
Configure in config.py:
| Setting | Default | Purpose |
|---|---|---|
MAX_ORDER_TOTAL_INR |
3000 | Hard price ceiling per order |
REQUIRE_CONFIRM_TOKEN |
True | Forces dry-run before every purchase |
CONFIRM_TOKEN_TTL_SECONDS |
300 | Token expiry (forces fresh price check) |
MCP Client Config
Example config file: mcp-config-example.json
Claude Desktop
%APPDATA%\Claude\claude_desktop_config.json (Windows) or ~/Library/Application Support/Claude/claude_desktop_config.json (macOS):
{
"mcpServers": {
"shop-mcp": {
"command": "C:\\path\\to\\shop-mcp\\.venv\\Scripts\\python.exe",
"args": ["C:\\path\\to\\shop-mcp\\server.py"],
"env": { "PYTHONUNBUFFERED": "1" }
}
}
}
Kiro IDE
Add to .kiro/settings/mcp.json:
{
"mcpServers": {
"shop-mcp": {
"command": "python",
"args": ["server.py"],
"cwd": "D:\\path\\to\\shop-mcp",
"env": { "PYTHONUNBUFFERED": "1" },
"disabled": false
}
}
}
Linux / macOS
{
"mcpServers": {
"shop-mcp": {
"command": "/path/to/shop-mcp/.venv/bin/python",
"args": ["/path/to/shop-mcp/server.py"],
"env": { "PYTHONUNBUFFERED": "1" }
}
}
}
Always use the Python from your
.venvso all dependencies are available.
Project Structure
shop-mcp/
├── server.py # MCP server — tool definitions & checkout guards
├── config.py # Safety settings, timeouts, site URLs
├── requirements.txt # Python dependencies
├── mcp-config-example.json # Client configuration template
├── browser/
│ ├── base.py # Playwright session management (cookies, context)
│ ├── amazon.py # Amazon.in browser actions (search, cart, checkout)
│ ├── flipkart.py # Flipkart browser actions
│ └── login.py # Manual login helper (opens browser for you)
└── sessions/ # Auto-created, .gitignored — stores login cookies
Important Notes
| Terms of Service | Amazon.in and Flipkart restrict bot access. This is for personal, low-volume use of your own account only. |
| Selectors Break | Both sites change their DOM frequently. Expect to update CSS selectors in browser/amazon.py and browser/flipkart.py when things stop working. |
| Headed Mode | Both sites detect headless browsers. Everything runs with a visible Chromium window (headless=False). |
| Session Expiry | When cookies expire, re-run python -m browser.login <site> to refresh. |
Tech Stack
| Component | Technology |
|---|---|
| Server Framework | MCP (Model Context Protocol) via fastmcp |
| Browser Automation | Playwright (Chromium) |
| Language | Python 3.10+ |
| Transport | stdio (local server) |
License
MIT License. See LICENSE for details.
<p align="center"> <i>Built for personal use. Use responsibly. Your account, your risk.</i> </p>
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.