AI & LLMs
Canopy API MCP
The Canopy API MCP server is a TypeScript implementation of the Model Context Protocol that gives AI assistants and agents direct access to Amazon product data. It runs on Cloudflare Workers and exposes 11 tools covering product lookup, search, categories, sellers, and more.
Installation
Clone and deploy the MCP server from GitHub:
git clone https://github.com/canopy-api/canopy-api-mcp
cd canopy-api-mcp
npm install
Start a local development server:
npm run dev
The server will be available at http://localhost:8787/mcp.
Deploy to Cloudflare Workers:
npm run deploy
Connecting to Claude Desktop
Add the server to your Claude Desktop configuration file (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):
{
"mcpServers": {
"canopy": {
"url": "http://localhost:8787/mcp",
"headers": {
"Authorization": "Bearer <YOUR_API_KEY>"
}
}
}
}
Replace <YOUR_API_KEY> with your key from the Canopy API dashboard. After saving, restart Claude Desktop — the Canopy tools will appear in the tools panel.
Authentication
The server accepts your API key in any of these header formats:
CANOPY-API-KEY: your-api-key
API-KEY: your-api-key
X-API-KEY: your-api-key
Authorization: Bearer your-api-key
Available Tools
The server exposes 11 tools that AI assistants can invoke:
| Tool | Description |
|---|---|
get_amazon_product | Product details by ASIN, URL, or GTIN |
get_amazon_product_variants | Product variants and options |
get_amazon_product_stock | Stock level estimates |
get_amazon_product_sales | Sales estimates (weekly/monthly/annual) |
get_amazon_product_reviews | Top customer reviews |
search_amazon_products | Search with filters and pagination |
get_amazon_autocomplete | Search suggestions |
get_amazon_categories | Root category taxonomy |
get_amazon_category | Category details with subcategories |
get_amazon_seller | Seller info and listings |
get_amazon_author | Author info and books |
Testing with MCP Inspector
You can test the server using the official MCP Inspector before connecting it to an AI client:
npx -y @modelcontextprotocol/inspector@latest
Connect the inspector to your running server URL (e.g. http://localhost:8787/mcp) and invoke tools interactively.