Hub Postman

Hub Postman is an MCP server that connects AI assistants directly to the Postman API platform, enabling natural language interaction with collections, environments, workspaces, and API workflows. This MCP bridge allows language models to browse API documentation, execute requests from collections, manage environment variables, and automate API testing workflows through conversational commands.

When to Use This MCP Server

Connect this server when...

• You want AI-assisted API exploration by browsing and executing requests from your Postman collections
• Your team maintains API documentation in Postman and you need conversational access to endpoint specifications
• You are building automated testing workflows that run Postman collection requests and validate responses
• You need to manage environment variables across multiple Postman environments for different deployment stages
• You want to generate API client code or documentation from your Postman collection definitions

Consider alternatives when...

• You only need to make simple HTTP requests without Postman collection management (use a generic fetch MCP server)
• Your API documentation lives in OpenAPI/Swagger files without Postman integration
• You need real-time API monitoring rather than on-demand request execution

Quick Start

Copy
# .mcp.json configuration
{
  "mcpServers": {
    "postman": {
      "type": "http",
      "url": "https://mcp.postman.com/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_POSTMAN_API_KEY"
      }
    }
  }
}

Connection setup:

1. Sign in to your Postman account at postman.co
2. Generate an API key from Account Settings > API Keys
3. Add the configuration above to your .mcp.json file with your API key
4. Restart your MCP client to connect to the Postman MCP endpoint

Example tool usage:

# Browse collections
> List all API collections in my default workspace

# Execute an API request
> Run the "Get User Profile" request from the Users API collection using the staging environment

# Manage environments
> Show me all environment variables in the production environment

Core Concepts

Concept	Purpose	Details
Collections	API request grouping	Organized sets of API requests with documentation, test scripts, and folder hierarchy
Environments	Variable management	Named sets of key-value pairs that parameterize API requests for different deployment stages
Workspaces	Collaboration scope	Team or personal containers that hold collections, environments, and shared API resources
HTTP Transport	Cloud connectivity	Connects to Postman's hosted MCP endpoint over HTTPS without requiring local server installation
API Key Auth	Secure access	Bearer token authentication using your Postman account API key for authorized platform access

Architecture:

+------------------+       +------------------+       +------------------+
|  Postman         |       |  Postman MCP     |       |  AI Assistant    |
|  Platform        |<----->|  Endpoint        |<----->|  (Claude, etc.)  |
|  (Cloud API)     | HTTPS |  (hosted HTTP)   | HTTP  |                  |
+------------------+       +------------------+       +------------------+
        |
        v
+--------------------------------------------------+
|  Workspaces > Collections > Requests > Envs       |
+--------------------------------------------------+

Configuration

Parameter	Type	Default	Description
Authorization	string	(required)	Bearer token using your Postman API key for authenticated access to the platform
url	string	https://mcp.postman.com/mcp	Postman MCP endpoint URL (use /minimal for reduced tool set)
workspace_id	string	(default workspace)	Specific workspace ID to scope operations to a particular team workspace
timeout	integer	30	Request timeout in seconds for API calls to the Postman platform
include_responses	boolean	true	Whether to include saved example responses when retrieving collection requests

Best Practices

1. Use workspace scoping for team environments. If you belong to multiple Postman workspaces, specify the workspace_id to ensure the AI assistant operates within the correct context. This prevents confusion between similarly named collections across personal and team workspaces.

2. Leverage environments for safe request execution. Always specify the appropriate environment (development, staging) when executing API requests through the MCP server. Never point AI-executed requests at production environments unless you have confirmed the request is safe and idempotent.

3. Organize collections with clear folder structures. Well-organized collections with descriptive folder names and request descriptions make it easier for the AI to find and execute the correct endpoints. Include documentation strings in your Postman collections to provide context for each API operation.

4. Use the minimal endpoint for focused workflows. The /minimal MCP endpoint provides a reduced set of tools optimized for common operations. Use this endpoint when you need faster response times and do not require the full range of Postman management capabilities.

5. Rotate API keys regularly and use team keys for shared access. Generate separate API keys for MCP server access rather than using your personal key for everything. Rotate keys quarterly and revoke any compromised keys immediately from your Postman account settings.

Common Issues

"401 Unauthorized" response from the MCP endpoint. Verify your Postman API key is valid and has not expired. Generate a new key from Account Settings > API Keys if needed. Ensure the Authorization header format is exactly "Bearer YOUR_KEY" with a space between "Bearer" and the key value.

Collections not appearing in query results. Check that you are querying the correct workspace. Collections in personal workspaces are not visible when the server is scoped to a team workspace and vice versa. Use the workspace listing tool first to identify available workspaces and their IDs.

Request execution fails with variable resolution errors. This occurs when a request references environment variables that are not defined in the selected environment. Ensure the environment contains all required variables before executing requests. Ask the AI to list environment variables and compare them with the request's variable references.