Project 1: Fashion Inventory MCP Cloud Agent

This project is deployed on Railway. It owns the inventory flat file, searches it, and uses OpenAI to interpret natural-language shopping requirements.

Important files

• cloud_agent.py: MCP server, inventory search tools, and cloud AI agent.
• data/inventory.json: 930 fictional fashion items across 31 categories.
• generate_inventory.py: reproducibly rebuilds the inventory file.
• railway.json: Railway start command.
• .env.example: required variable names without real secrets.

Cloud AI role

The cloud AI translates language such as "black waterproof-style jacket in medium under 150 euros" into arguments for search_inventory. Python—not AI—filters the flat file and returns exact products, prices, sizes, and stock. The AI then selects and explains matches without inventing inventory.

It also exposes recommend_matching_items. For a question such as "what trousers go with a navy shirt?", the cloud AI obtains approved color pairings, searches real trouser stock, and returns only available SKUs.

Run locally on Mac

python3 -m venv .venv
source .venv/bin/activate
python -m pip install -r requirements.txt
cp .env.example .env

Add real values to .env, then:

python cloud_agent.py

The local test endpoint is http://127.0.0.1:8000/mcp.

Deploy on Railway

Push the parent repository to GitHub and create a Railway service. Because this is a subfolder, configure:

Root Directory: /fashion-inventory-system/cloud-inventory-mcp
Config File Path: /fashion-inventory-system/cloud-inventory-mcp/railway.json

Add Railway variables:

MCP_SECRET_TOKEN=your-generated-shared-secret
OPENAI_API_KEY=your-cloud-openai-key
CLOUD_OPENAI_MODEL=gpt-4.1-mini

Do not define PORT; Railway supplies it. Generate a public domain. The MCP endpoint will be https://YOUR-SERVICE.up.railway.app/mcp.

See the parent OVERVIEW_AND_SETUP.md for the complete two-project walkthrough.

git status git add cloud_agent.py generate_inventory.py data/inventory.json README.md git commit -m "Add complete catalog and outfit recommendations" git push origin main