Amazon Business Integrations MCP Server
The Amazon Business Integrations Model Context Protocol (MCP) Server empowers AI-enabled developer environments by providing direct, contextual access to Amazon Business API documentation, sample code, and troubleshooting resources. This server acts as a bridge between your IDE and Amazon Business’s technical knowledge base, enabling AI models to instantly retrieve relevant API details, generate sample integration code, and offer guided solutions for common errors. You can use this MCP Server throughout your API integration process to simplify and accelerate development.
This MCP Server is currently in preview release and is intended for testing and evaluation purposes. This MCP Server is subject to Amazon’s Conditions of Use.
Features
- Search Documentation: Search through documentation using natural language queries.
- Read Documentation: Retrieve documentation content using references.
With these functions, you can search for and read Amazon Business API documentation to understanding various API offerings, and extract sample code during your integration.
Prerequisites
Before using the Amazon Business Integrations MCP Server, complete these steps:
- Ensure that you have installed an IDE that supports MCPs and developer agents, such as Amazon Q.
- Install node.js and npm if you do not have them on your system.
- Access your account ID in your Solution Provider Portal (SPP) account. You can find this ID from the Global Accounts section in SPP.
You will use this ID to complete the MCP Server configuration. If you do not have this ID, complete the API onboarding process to generate your ID. For more information, see Onboarding overview.
If you only plan to use the MCP Server to search and read Amazon Business documentation, complete Onboarding Step 1 to create your developer account and generate your account ID. If you plan to use the MCP Server to help generate code, complete Onboarding Step 1, Step 2, and Step 3 to receive the access and refresh tokens required for production API calls.
Installation
Follow the appropriate steps below to install the MCP server in your preferred IDE:
- Add Amazon Business Integrations MCP Server to the MCP configuration file (for example, for Amazon Q Developer CLI, edit
~/.aws/amazonq/mcp.jsonby using the below JSON snippet). - In the
--dev-account-id=<account_id>field in theargslist, provide your account ID from Prerequisites Step 3:
{
"mcpServers": {
"ab-integrations-mcp-server": {
"command": "npx",
"args": ["@amazonbusiness/ab-integrations-mcp-server@latest", "--dev-account-id=<account_id>"],
"disabled": false,
"autoApprove": []
}
}
}
Quick start
Once you’ve installed the Amazon Business Integrations MCP Server, you can ask your AI assistant questions such as:
- “I want to build a react-based shopping application that displays products and product detail pages using Product Search API.”
- “I want to build a shopping application that displays products using Product Search API and places orders using Ordering API.”
- "Search for information about the Product Search API"
- "Find documentation about authentication requirements"
- “How can I test my Amazon Business API calls?”
Available MCP tools
search-documentation
Perform semantic searches across the documentation corpus using natural language queries. Returns relevant documentation snippets with document references.
Here is the folder structure that determines the documentationPath to use for searches in the MCP server:
📂
├── API Swagger Models/ - API schemas, data structures, request/response formats, field definitions, and models for building applications and generating code
├── OAuth/ - Essential authentication and authorization documentation covering tokens, login processes, credentials, access control, and security implementation. OAuth authentication is a prerequisite for all Amazon Business API implementations.
├── Ordering API/ - Order placement, creation, purchasing, and buying operations
├── Product Search API/ - Product searching, catalog browsing, item discovery, and product details
├── Reconciliation API/ - Purchase data reconciliation, transaction matching, and financial records management
├── Reporting API/ - Order history, tracking information, payment details, and analytics reports
├── User Management API/ - User account management and user operations
├── Document API/ - Invoice generation, receipts, and document reports
├── Application Management API/ - Client secrets, application management, and secret rotation
└── Guides and FAQs/ - Step-by-step tutorials, implementation guides, frequently asked questions, best practices, getting started resources, and practical examples
Parameters
| Name | Description | Schema | Required |
|---|---|---|---|
| query | Natural language search query to find documentation. | string | Yes |
| maxResults | Maximum number of results to return. (default: 5, maximum: 20) | number | No |
| documentationPath | List of documentation paths to search (e.g. API Swagger Models, OAuth, Ordering API, Product Search API) to narrow down the search results (maximum: 5). | string | No |
read-documentation
Retrieve complete documentation content using document references obtained from documentation and search results.
Parameters
| Name | Description | Schema | Required |
|---|---|---|---|
| documentReference | Reference to the document from the search results. | string | Yes |
Authentication
The Amazon Business Integrations MCP Server does not require authentication but is subject to rate limits.
For support or questions about the MCP Server, please contact abapi-integrations@amazon.com.
Disclaimer
This MCP Server is provided as an open-source tool to allow users to access their own AI tools to query documentation and generate sample code. The sample code generated is intended solely for demonstration and educational purposes. Amazon Business makes no representations and bears no responsibility related to the accuracy, completeness, or suitability of any generated code. Amazon’s Conditions of Use apply.
Users are advised to exercise their own discretion and perform independent verification before using any code in production or critical environments. The maintainers and contributors of this project shall not be held liable for any errors, omissions, or damages arising from the use of this server or its outputs.
By using the MCP Server, you acknowledge that you are responsible for evaluating the suitability and correctness of the generated content for your specific use case.