MCP Api Client

🚀 mcp-api-client

Postman for LLMs. A specialized, resource - efficient MCP server designed for rapid REST API testing and exploration.

This is a specialized, resource - efficient MCP server crafted for rapid REST API testing and exploration. Instead of writing boilerplate test code, your LLM assistant can now directly interact with, debug, and validate your endpoints in real - time. It is built with FastMCP and httpx.

🚀 Quick Start

Traditional API testing often involves writing scripts or switching to GUI tools like Postman. The mcp-api-client server integrates the power of API testing directly into your AI - assisted development workflow.

✨ Features

• LLM - Native Testing: Quickly test endpoints using natural language without writing test code.
• Resource Efficient: The minimal response format reduces token consumption, only returning status and body by default.
• Zero - Boilerplate Prototyping: Interact with new APIs instantly to understand their behavior.
• Local Dev Companion: Ideal for testing your local development server (localhost) during the build phase.

📦 Installation

Requirements: Python 3.10+

git clone https://github.com/etaaa/mcp-api-client.git
cd mcp-api-client
pip install .

📚 Documentation

Setup

1. Find the installed path:

   which mcp-api-client
   

2. Add the server to your MCP client configuration (e.g., Claude Desktop):

   {
     "mcpServers": {
       "api-client": {
         "command": "/full/path/to/mcp-api-client"
       }
     }
   }
   

Tools

http_request

Make a single HTTP request with minimal response data.

Response format:

{
  "status": 200,
  "body": {"key": "value"}
}

http_batch_request

Execute multiple HTTP requests efficiently in a single call. Highly recommended for testing multiple endpoints to minimize token usage.

Example:

{
  "requests": [
    {"method": "GET", "url": "http://localhost:8000/health"},
    {"method": "POST", "url": "http://localhost:8000/users", "body": {"name": "Test"}},
    {"method": "GET", "url": "http://localhost:8000/users/1"}
  ]
}

Returns: Array of responses in the same order as requests.

💻 Usage Examples

Basic Usage

Single request:

method: GET
url: http://localhost:8000/health

Advanced Usage

Batch testing (recommended for multiple endpoints):

requests:
  - method: GET
    url: http://localhost:3000/api/health
  - method: POST
    url: http://localhost:3000/api/users
    body: {"name": "Test User", "role": "admin"}
  - method: GET
    url: http://localhost:3000/api/users

Agent Skills

This project includes a pre - written "skill" that instructs LLM agents on how to best use this tool.

Usage:

1. Copy skills/mcp-api.client.md to your project's agent configuration (e.g., .cursor/skills/ or similar).
2. Your agent will now understand when to use this server (e.g., for verifying API changes).

Testing & Contributing

pip install pytest pytest-asyncio
pytest

Contributions are welcome! If you find a bug or have a feature request, please open an issue or submit a PR.

📄 License

This project is licensed under the MIT License. See the LICENSE file for details.