Standalone MCP server that lets AI assistants send, save and test HTTP requests (Postman-like tooling for MCP).
https://github.com/hijaz/postmancerStop switching between Claude and Postman. Postmancer brings full REST client functionality directly into your AI assistant conversations, letting you test, debug, and document APIs without breaking your flow.
You're deep in a coding session with Claude, working through an API integration. The conversation is flowing, solutions are emerging, then you hit the inevitable: "Let me test this endpoint real quick."
Out comes Postman. You craft the request, test it, screenshot the response, paste it back into Claude. Flow broken. Context lost. This dance repeats dozens of times per project.
Postmancer eliminates this friction by giving Claude (or any MCP-compatible AI) the ability to make HTTP requests, save collections, manage environments, and analyze responses - all within your conversation.
What you get:
API Integration Debugging: Instead of describing error responses to Claude, it sees them firsthand. It can suggest fixes based on actual response codes, headers, and payloads.
Documentation Generation: Claude can test your entire API surface, then generate accurate docs based on real request/response patterns rather than guessing from code.
Environment Verification: Deploy to staging? Claude can validate your endpoints work as expected across different environments by running your saved collections.
Integration Testing: Build test suites conversationally. "Test the user creation flow" becomes a natural language command that executes a sequence of authenticated requests.
Install and configure in under two minutes:
npm install -g postmancer
Add to your Claude Desktop config:
{
"mcpServers": {
"postmancer": {
"command": "npx",
"args": ["-y", "postmancer"]
}
}
}
Now Claude can make requests like:
"Test the user registration endpoint with this payload and show me the response structure" - Claude executes the request, analyzes the response, and suggests improvements or next steps.
Built with security in mind. Environment variables and credentials are encrypted locally. Your API keys never leave your machine, and you can use custom encryption keys in production environments.
The tool provides seven core capabilities through natural language:
This isn't about replacing your REST client - it's about making your AI assistant fluent in HTTP so you never have to leave the conversation to test an idea.
Your API development workflow just got conversational.