Pangolinfo Amazon Data MCP

Pangolinfo Amazon Data MCP gives AI agents real-time access to Amazon commerce intelligence, WIPO design-patent data, Google Trends and search intelligence, and local market data.

This package is Pangolinfo's official stdio bridge to the hosted Streamable HTTP MCP server. It discovers the current remote tool schemas at startup and forwards MCP tool calls without storing your API key or result data.

Capabilities

The server exposes 19 business data tools, plus the free `pangolinfo_capabilities` navigation tool:

• Amazon research: product search and detail, reviews, delivery estimates, best sellers, new releases, seller catalogs, category products, and Rufus recommendations.
• Market selection: category discovery and trees, category metrics, niche filtering, and category paths.
• Broader intelligence: WIPO design-patent and litigation-risk search, Google SERP/AI Overview search, Google Trends, Google Maps local search, and a generic Amazon URL scraper.

See the full product overview and current capability details at pangolinfo.com/amazon-data-mcp.

Requirements

• Node.js 18 or newer
• A Pangolinfo API key

Run with npx

PANGOLINFO_API_KEY="your-key" npx -y pangolinfo-amazon-data-mcp

Do not put a production key in source code or commit it to Git.

MCP client configuration

For clients that launch local stdio servers:

{
  "mcpServers": {
    "pangolinfo-amazon-data": {
      "command": "npx",
      "args": ["-y", "pangolinfo-amazon-data-mcp"],
      "env": {
        "PANGOLINFO_API_KEY": "your-key"
      }
    }
  }
}

Clients with native Streamable HTTP support can connect directly:

• Endpoint: `https://mcp.pangolinfo.com/mcp`
• Header: `Authorization: Bearer YOUR_PANGOLINFO_API_KEY`

Agent-oriented installation instructions are also available in `llms-install.md`.

Run with Docker

Pull the public multi-platform image from Docker Hub or GitHub Container Registry:

docker pull pangolinfo/amazon-data-mcp:latest
docker pull ghcr.io/pangolin-spg/amazon-data-mcp:latest

Run it as a stdio MCP server while passing the key only at runtime:

docker run --rm -i \
  -e PANGOLINFO_API_KEY \
  pangolinfo/amazon-data-mcp:latest

Or build the official bridge locally:

docker build -t pangolinfo-amazon-data-mcp .

Run the local image:

docker run --rm -i -e PANGOLINFO_API_KEY pangolinfo-amazon-data-mcp

Set `PANGOLINFO_API_KEY` in your shell or secret manager. Do not bake it into the image, Dockerfile, or source tree.

How the bridge works

The package opens an authenticated Streamable HTTP connection to Pangolinfo, then exposes the remote `tools/list` and `tools/call` methods over local stdio. Tool definitions therefore stay synchronized with the hosted service; the package contains no embedded customer data and does not log credentials.

An optional `PANGOLINFO_MCP_URL` environment variable can override the endpoint for approved testing environments. For safety, only HTTPS URLs are accepted.

Links

• Amazon Data MCP product page
• Amazon Scraper API
• AI Overview SERP API
• Amazon Niche Data API
• Amazon Alexa API
• Pangolinfo website
• GitHub repository
• Docker Hub image
• Canonical MCP Registry source
• Issue tracker
• MCP client setup guide
• Amazon API documentation
• AI Overview API documentation
• Amazon Niche API documentation
• Amazon Alexa API documentation

License and trademarks

The bridge source code is licensed under the MIT License. Pangolinfo names, logos, product marks, and brand assets are not granted under that license; see BRANDING.md.