GWP Custom MCP Server

01 · The problem

The Global Wholesale Products platform had mature Flask APIs over an Elasticsearch product database, and the team wanted users to query and manage it conversationally through Anthropic's Claude.

02 · What I built

A custom MCP server in Python with FastMCP: 20 tools mapped to 18 existing APIs — product search, pricing lookups, Excel report generation, and record modification — each with parameter schemas and semantic descriptions so Claude could use them autonomously. I translated the team's entire Postman collection into a production tool set.

03 · The impact

Built and deployed to production in a single day — the second-to-last day of my internship — as an early adopter of MCP before widespread industry uptake. Integrated for logged-in users at globalwholesaleparts.com.

What I did

• Built a 20-tool MCP server mapped to 18 existing APIs using FastMCP in Python.
• Defined tool schemas, parameters, and semantic descriptions for search, pricing, reporting, and data-modification APIs.
• Onboarded onto an unfamiliar project, translated a full Postman collection into production MCP tools, and deployed same-day.
• Early adoption of Anthropic's Model Context Protocol before widespread industry uptake.