You will get a working MCP server that connects Claude to your API or data source
Project details
I will build an MCP (Model Context Protocol) server that connects Claude or any AI assistant directly to your API or data source.
What MCP means for you: instead of copy-pasting data into Claude, you just ask naturally - "look up this record", "check this status", "summarize today's data" - and Claude fetches the real answer live from your system.
What you get:
• Python MCP server with clearly defined tools
• Tool descriptions written so Claude knows exactly when and how to use them
• Async API client with retry logic and proper error handling
• Unit tests with mocked network calls
• Setup instructions for Claude Desktop and Claude Code
• Real demo showing Claude calling the tools live
I've built an MCP server wrapping the Romanian ANAF public API — 3 tools, 19 passing tests, real screenshots of Claude using it live. You can review the code and quality before ordering.
Not sure if MCP fits your use case? Message me first and I'll tell you
honestly.
What MCP means for you: instead of copy-pasting data into Claude, you just ask naturally - "look up this record", "check this status", "summarize today's data" - and Claude fetches the real answer live from your system.
What you get:
• Python MCP server with clearly defined tools
• Tool descriptions written so Claude knows exactly when and how to use them
• Async API client with retry logic and proper error handling
• Unit tests with mocked network calls
• Setup instructions for Claude Desktop and Claude Code
• Real demo showing Claude calling the tools live
I've built an MCP server wrapping the Romanian ANAF public API — 3 tools, 19 passing tests, real screenshots of Claude using it live. You can review the code and quality before ordering.
Not sure if MCP fits your use case? Message me first and I'll tell you
honestly.
AI Development Type
Knowledge Representation, Recommendation SystemAI Development Language
PythonWhat's included $35
These options are included with the project scope.
$35
- Delivery Time 5 days
- Number of Revisions 2
Frequently asked questions
About Catalin
Python Automation Developer - n8n Nodes, MCP Servers, API Integrations
Bucharest, Romania - 1:27 am local time
Recent work (all public on GitHub):
- 5 published n8n community nodes on npm — connecting Romanian business data
sources (ANAF, BNR, SEAP, BVB, imobiliare.ro) to n8n workflows
- MCP server wrapping the ANAF public API — lets Claude query official Romanian
company records live, with 3 tools and 19 passing tests
- Python CLI pipeline — batch ANAF company lookups → color-coded Excel report,
async httpx, pydantic validation, re-runnable with SHA256 change detection
Day job: Configuration Engineer at FlowX.AI, where I work on enterprise process automation, Python/JavaScript scripting, and REST API integrations for large clients.
What I take on freelance:
→ Custom n8n nodes for APIs that don't have community nodes yet
→ MCP servers connecting Claude/AI assistants to your data sources
→ Python automation scripts — API integrations, data pipelines, Google Sheets
automation, web scraping
→ Small REST APIs (FastAPI/Flask) with clean documentation
I work async, fixed-price, and I deliver clean code with proper READMEs so you can maintain it without me. No surprises on scope.
Steps for completing your project
After purchasing the project, send requirements so Catalin can start the project.
Delivery time starts when Catalin receives requirements from you.
Catalin works on your project following the steps below.
Revisions may occur after the delivery date.
I review the API and confirm which tools to expose within 24 hours.
I build the MCP server, tools, and async API client with error handling.