Every business is racing to make data usable by AI assistants like ChatGPT, Claude, or Perplexity. The challenge? Your data sits behind APIs too complex for non-developers. Training an LLM to speak your API’s language is costly and unreliable. That’s where MCP servers (Model Context Protocol) step in.
This repo shows how to take a GraphQL schema and turn it into a fully functional MCP server using Ariadne Codegen and FastMCP. The result: structured, type-safe tools your AI assistant can call directly.
- APIs are too complex for non-developers.
- Training LLMs to use APIs is error-prone and expensive.
- MCP servers bridge the gap: you expose tools, not docs.
Example query from ChatGPT → API call:
"Show me women’s shoes under 100 PLN available in the Polish channel"
This becomes a validated GraphQL query executed against your backend, no custom UI required.
MCP servers are the new extension surface for AI assistants. They let you expose your existing APIs as tools:
- No plugin marketplace friction.
- No extra UI.
- Just direct access for AI.
For e-commerce, SaaS, or any data-driven business, this means product discovery, catalog browsing, and customer support become instantly AI-ready.
- GraphQL – schema-first, typed APIs
- Ariadne Codegen – generates async GraphQL clients + Pydantic models
- FastMCP – framework for MCP server hosting & tool exposure
git clone https://github.com/mirumee/mcp-ariadne-codegen-example
cd mcp-ariadne-codegen-example
python -m venv .venv
source .venv/bin/activate
pip install -e .
mcp
Postman now supports MCP requests natively. You can call endpoints, inspect request/response, and validate before exposing them to ChatGPT. See Postman MCP docs for details.
We implemented this demo using Saleor, a headless e-commerce platform with a rich GraphQL API.
Focus area: the products query — complex filters, search, ordering, and cursor pagination.
Instead of teaching ChatGPT cursor-based pagination, we wrapped Saleor GraphQL with Ariadne Codegen + FastMCP. The assistant can now call typed tools directly.
Demo storefront: nimara.store.
A special shout-out goes to @maarcingebala from the Saleor team. He first suggested the idea of combining GraphQL code generation with MCP, and even published a reference project at saleor-mcp . His work laid the foundation for this demo and shows how far we can push GraphQL when paired with the new MCP ecosystem.