MCP Product Management System

# Sample Questions for MCP Product Management System # These questions demonstrate the various capabilities of your enhanced MCP server # Use these with Claude Desktop, MCP Inspector, or the Ollama Bridge ## Basic Product Queries 1. What are the details of SKU 120715? 2. Show me the first 5 products in our inventory 3. Give me details about product 2420981 4. What's the information for item 1383264? ## Search & Discovery 5. Find all products with "shirt" in the name 6. Search for blue products in our inventory 7. Show me all polo shirts we have 8. Find products from American Apparel 9. Search for products containing "cotton" 10. What electronics do we carry? ## Category & Classification 11. What product categories do we have? 12. Show me all products in the T-Shirts category 13. List products in the Sweatshirts/Fleece category 14. What subcategories exist under our main categories? ## Stock Management 15. Which products have low stock (under 50 units)? 16. Show me products that are completely out of stock 17. Find products with more than 1000 units in stock 18. What's our stock situation for products under 10 units? ## Django-Style Advanced Filtering 19. Filter products where category contains "shirt" and stock is greater than or equal to 100 20. Show me size XL products with prices less than $20, sorted by stock descending 21. Find all black products in sizes S, M, and L from the Gildan warehouse 22. Filter products with prices between $10 and $25, sorted by price ascending 23. Show me page 2 of T-Shirts with more than 50 units in stock, 10 items per page ## Analytics & Statistics 24. What are the filter statistics for categories, colors, and sizes? 25. Give me a breakdown of our inventory by warehouse and product status ## Complex Business Queries (for Ollama Bridge) These work especially well with the Ollama natural language processing: - "I need to find casual shirts under $15 that we have good stock levels for" - "What blue items do we have that are popular and well-stocked?" - "Show me our most expensive products that are running low" - "Find professional-looking items suitable for corporate gifts" - "What products would work well for a summer promotional campaign?" ## Sample MCP Tool Calls (JSON format for direct testing) ### Basic Product Lookup: ```json { "tool": "get_product", "arguments": {"sku": "120715"} } ``` ### Django-Style Filtering: ```json { "tool": "filter_products", "arguments": { "filters": { "category__icontains": "shirt", "stock__gte": 50, "price__lte": 25.00 }, "ordering": ["-stock", "title"], "page": 1, "page_size": 10 } } ``` ### Advanced Search: ```json { "tool": "advanced_search_products", "arguments": { "query": "blue", "min_stock": 20, "category_filter": "T-Shirts", "sort_by": "price", "limit": 15 } } ``` ### Filter Statistics: ```json { "tool": "get_filter_stats", "arguments": { "fields": ["category", "color", "size", "warehouse", "status"] } } ``` ### Size-Specific Filtering: ```json { "tool": "filter_products", "arguments": { "filters": { "size__in": ["S", "M", "L"], "color__icontains": "black", "warehouse__exact": "Gildan" }, "ordering": ["price"], "page": 1, "page_size": 20 } } ``` ## Tips for Best Results: 1. **Natural Language (Ollama)**: Ask conversational questions - the system understands context 2. **Specific Filtering**: Use Django-style filters for precise results 3. **Pagination**: For large result sets, use page and page_size parameters 4. **Multiple Criteria**: Combine multiple filters for targeted searches 5. **Sorting**: Use ordering arrays with "-" prefix for descending sorts ## Testing Different Interfaces: - **Claude Desktop**: Use natural language questions (1-25) - **MCP Inspector**: Test JSON tool calls directly - **Ollama Bridge**: Try conversational queries with different models - **API Testing**: Use the JSON examples for programmatic access ## Performance Notes: - Filter queries are optimized with proper indexing - Pagination prevents memory issues with large datasets - Statistics queries provide insights into data distribution - Advanced search supports complex multi-field queries Happy testing! 🚀