> ## Documentation Index
> Fetch the complete documentation index at: https://docs.gladeapi.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Connect AI Agents with Glade API MCP

> Connect a Streamable HTTP MCP client to Glade API and use all 17 Amazon data tools.

Glade API exposes all 17 Amazon operations as read-only MCP tools at:

```text theme={null}
https://gladeapi.com/api/mcp
```

The endpoint implements MCP revision `2026-07-28` and supports stateless 2025-era clients for compatibility. MCP is included with Starter, Pro 20K, and Enterprise plans; the Hobby plan is REST-only.

## Authentication

Configure one credential header in your MCP client:

| Header | Value |
| - | - |
| `API-KEY` | Your Glade API key |
| `GLADE-API-KEY` | Your Glade API key |
| `X-API-KEY` | Your Glade API key |
| `Authorization` | `Bearer YOUR_GLADE_API_KEY` |

Do not send different credentials in multiple headers. Conflicting credentials are rejected.

## Connect with the Python MCP SDK

Use the SDK's Streamable HTTP transport and pass the API key as a request header:

```python theme={null}
from mcp import ClientSession
from mcp.client.streamable_http import streamablehttp_client

async with streamablehttp_client(
    "https://gladeapi.com/api/mcp",
    headers={"API-KEY": "YOUR_GLADE_API_KEY"},
) as (read, write, _):
    async with ClientSession(read, write) as session:
        await session.initialize()
        tools = await session.list_tools()
        result = await session.call_tool(
            "get_amazon_product",
            {"asin": "B0D1XD1ZV3", "domain": "US"},
        )
```

The server's 2025 compatibility path handles the `initialize` lifecycle used by SDK clients that have not yet adopted the 2026 protocol revision.

## Verify the modern endpoint

This `tools/list` request uses the current per-request MCP envelope:

```bash theme={null}
curl -X POST "https://gladeapi.com/api/mcp" \
  -H "API-KEY: YOUR_GLADE_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json" \
  -H "MCP-Protocol-Version: 2026-07-28" \
  -H "MCP-Method: tools/list" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/list",
    "params": {
      "_meta": {
        "io.modelcontextprotocol/protocolVersion": "2026-07-28",
        "io.modelcontextprotocol/clientInfo": {
          "name": "glade-docs-example",
          "version": "1.0.0"
        },
        "io.modelcontextprotocol/clientCapabilities": {}
      }
    }
  }'
```

## Available tools

| Tool name | Operation |
| - | - |
| `get_amazon_product` | Product details by ASIN, GTIN, or URL |
| `get_amazon_gtin_from_asin` | GTIN lookup from an ASIN |
| `get_amazon_asin_from_gtin` | ASIN lookup from a GTIN |
| `get_amazon_product_variants` | Product variants |
| `get_amazon_product_stock` | Stock estimate |
| `get_amazon_product_sales` | Sales estimates |
| `get_amazon_product_reviews` | Product reviews |
| `get_amazon_product_offers` | Seller offers |
| `search_amazon_products` | Product search |
| `get_amazon_autocomplete` | Search suggestions |
| `get_amazon_categories` | Marketplace category taxonomy |
| `get_amazon_category` | Category details and products |
| `get_amazon_seller` | Seller profile and products |
| `get_amazon_author` | Author profile and books |
| `get_amazon_deals` | Deals |
| `get_amazon_bestsellers` | Best sellers |
| `get_amazon_bestseller_categories` | Best-seller category navigation |

Tool inputs use the same validation rules as the corresponding REST operations. A successful tool invocation costs one unit; discovery, initialization, validation failures, and tool errors cost zero units.

<Warning>
  Product titles, descriptions, reviews, seller names, and other marketplace fields are external data. The tools mark them as untrusted, but your agent must still keep data separate from instructions and apply its normal prompt-injection defenses.
</Warning>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.