MCP API Client (Beta)

A Model Context Protocol (MCP) server that call APIs using YAML config files.

This server allows you connect to APIs using simple a config YAML file! 🤩

Ideal for: #lazy-programmers #nocode #lowcode #ai-integrators #automation #api-gateway #data-pipelines #rapid-prototyping #llm-integration #makers

If you don't have advance APIs call skills using a programming language or you just want a fast way to connect your AI model with an API using MCP, you will LOVE 💖 this MCP.

🎬▶️ Example Video😉✨

Check this video using included Test Server and Claude

Related MCP server: OpenAI MCP Server

Support with your Click!

Make a reaction and give me a motivational comment on mcp.so repository

... and yes! A donation to help cover my coffee and rent would be amazing! 🙏🙏

Features

• Configure multiple API endpoints using YAML

• Automatic MCP tool creation from YAML config file.

• Automatic MCP fetch API creation from YAML config file.

• Supports all HTTP methods (GET, POST, PATCH, PUT, DELETE) [✨PARTIAL✨]

  • [✅] GET API without parameters

  • [✅] GET API with query parameters

  • [✅] GET API with path parameters

  • [✅] POST API

  • [✅] PATCH API with path parameters

  • [✅] PUT API with

  • [✅] DELETE API with path parameters

• Set API Headers [✨WORKING✨]

• Set API Token [✨ 🫣🔴COMMING SOON✨]

• Environment variable interpolation for sensitive data [✨ 🫣🔴COMMING SOON✨]

• Automatic PATH parameter detection from URL (like /users/{id})[✨WORKING✨]

• Automatic QUERY parameter detection from URL (like /users?page={page}&limit={limit}) [✨WORKING✨]

• Reduce Prompt tokens used setting up certain APIs for your project

• ...and best of all, SAVE COUNTLESS HOURS FOR HUMANITY 👏😎😉 by connecting your LLM to any API.

Use cases

• Connect to N8N, Make.com or Zapier webhook: Make actions in other platforms connecting to automations or integration tools.

• AI Model Integration: Quickly connect LLMs or other AI models to external APIs for data retrieval, enrichment, or action execution without custom code.

• Rapid Prototyping: Build and test API workflows using YAML configuration, ideal for hackathons, demos, or proof-of-concept projects.

• No-Code/Low-Code Automation: Enable non-developers to define and call APIs securely using simple YAML files and environment variables.

• Data Pipelines: Orchestrate API calls as part of ETL or data processing pipelines, leveraging MCP tools for modularity and reuse.

• API Gateway: Expose multiple backend APIs through a unified MCP interface, simplifying integration for downstream consumers.

• Secure API Access: Manage secrets and tokens via environment variables, keeping sensitive data out of source code and YAML files.

Getting started

1. Get a copy of default api.yaml.

npx -y mcp-api-client --init

2. You can test default api.yaml with express test server test/server.js. Execute in other terminal.

npx -y mcp-api-client --test-server

3. Add MCP to your agent config file.

• Example for Claude Client settings file:

"mcp": {
    "servers": {
      "mcp-api-client": {
          "command": "npx",
          "args": [
              "-y",
              "mcp-api-client",
              "path-to-api.yaml"
          ]
      },
    }
  }

• Example for Visual Studio Code settings file:

"mcp": {
    "servers": {
      "mcp-api-client": {
          "command": "npx",
          "args": [
              "-y",
              "mcp-api-client",
              "path-to-api.yaml"
          ]
      },
    }
  }

4. Just Ask!

YAML Configuration File

Each API YAML file can include global metadata with MCP information:

metadata: General information about the API set

and a list of API endpoint definitions:

apis: A list of API endpoint definitions. Each endpoint should have

Example API YAML config file example

metadata:
  name: API Examples
  version: 1.0.0
  description: This file contains examples of various API endpoints for testing purposes.

apis:
  - name: getUser
    url: https://api.example.com/users/{id}
    method: GET
    content:
      headers:
        Accept: application/json
  
  - name: createUser
    url: https://api.example.com/users
    method: POST
    api-token: ${API_TOKEN}
    content:
      headers:
        Content-Type: application/json
      body:
        name: string
        email: string
  ...

See public/apis.yaml for a complete example with metadata and all HTTP methods and parameter types.

API_TOKEN from .env (Not tested yet!)

Create a .env file for your API tokens:

API_TOKEN=your_api_token_here

Bugs and Requests

https://github.com/molavec/mcp-api-client/issues

Contributing

1. Fork the repository

2. Create your feature branch

3. Commit your changes

4. Push to the branch

5. Create a new Pull Request

License

MIT