mcp -postman-tool-generation

This MCP server exposes a single tool named generate_ai_tool that converts a Postman collection and request into a type-safe AI tool implementation. It integrates with the Postman API to fetch and process the specified collection and request, then outputs code compatible with multiple AI frameworks (OpenAI, Mistral, Gemini, Anthropic, LangChain, AutoGen) and languages (JavaScript/TypeScript). To use it, first ensure your MCP client can reach the server name postman-ai-tools, then invoke the tool with the required parameters: collectionId, requestId, language, and agentFramework. The generated code includes request/response type definitions, error handling, API integration stubs, and OpenAI function definitions, along with accompanying documentation and usage examples. When you run the tool, you will receive a ready-to-use code package that you can integrate into your applications or deploy as part of your AI tools library.

Prerequisites:

• Node.js and npm installed on your machine
• Access to a Postman API key
• MCP client or Claude app configured to reach your MCP server (postman-ai-tools)

Step-by-step installation:

# 1. Clone the server repository (or obtain the package)
# Example: git clone https://github.com/your-org/mcp-server-postman-tool-generation.git
cd mcp-server-postman-tool-generation

# 2. Install dependencies
npm install

# 3. Build the server (if required)
npm run build

# 4. Set the Postman API key in your environment or MCP settings
export POSTMAN_API_KEY=your-postman-api-key

Configuration in Claude or MCP client:

{
  "mcpServers": {
    "postman-ai-tools": {
      "command": "node",
      "args": [
        "/path/to/postman-tool-generation-server/build/index.js"
      ],
      "env": {
        "POSTMAN_API_KEY": "your-postman-api-key"
      },
      "disabled": false,
      "autoApprove": []
    }
  }
}

Usage:

• Ensure the server is running and reachable as postman-ai-tools.
• Use the MCP client to call generate_ai_tool with collectionId, requestId, language, and agentFramework.

Notes:

• The server relies on a valid Postman API key; keep it secret and rotate it periodically.
• The generated code targets multiple AI frameworks; choose the appropriate agentFramework at invocation time.
• If you encounter API rate limits from Postman, consider increasing retries or introducing caching at the MCP level.
• Ensure your build/index.js path in the mcp_config matches your deployment environment.
• For debugging, check server logs for errors related to missing collectionId or requestId, invalid IDs, or authentication failures.