Neolibrarian MCP

# Claude Desktop Test Prompt for Calibre Library MCP Server Use this prompt in Claude Desktop to thoroughly test your Calibre Library MCP server with your 107k+ book collection. --- ## Test Prompt for Claude Desktop ``` I want to thoroughly test my Calibre Library MCP server that has 100,000+ books. Please help me validate all functionality and performance. Run through these tests systematically and report on what works, what doesn't, and any performance issues: ## 1. Basic Connectivity & Health Check - Test the connection to my Calibre library - Get total book count and library statistics - Check if full-text search is available ## 2. Search Functionality Tests - Search for books by author "Stephen King" (should return many results) - Search for books with title containing "Lord of the Rings" - Try a complex search combining author "Tolkien" with format "EPUB" - Test fuzzy search with a misspelled author name like "Tolken" - Search for books published after 2020 ## 3. Batch Operations & Performance - Get details for 50 random books in a single batch request - Search for fantasy books and paginate through results (test offset/limit) - Try to get details for books with IDs [1, 100, 1000, 5000, 10000] ## 4. Content Extraction Tests - Get the content of a well-known book (try book ID 1 first) - Extract a sample from the beginning of a book - Try content extraction from different formats (EPUB, PDF, TXT if available) ## 5. Discovery & Browsing - Get 20 random books for discovery - Find books in a popular series - Get recent additions to the library ## 6. Full-Text Search (if available) - Search for specific text content within books - Try searching for a famous quote or phrase across all books ## 7. Edge Cases & Error Handling - Try to get details for a non-existent book (ID 999999) - Search with empty query strings - Request extremely large batch sizes (test limits) - Try invalid parameters ## 8. Performance Analysis For each operation, please note: - Response time (fast/slow/very slow) - Result accuracy and completeness - Any errors or warnings - Memory usage if noticeable ## 9. Unified Search Testing Test the new enhanced search endpoint with complex queries: - Multi-filter search (author + format + date range) - Sort by different criteria (title, date, relevance) - Large result sets with pagination ## Expected Behavior: - Sub-second responses for most operations - Accurate search results - Proper pagination with correct total counts - Graceful error handling - Consistent response formats Please run these tests systematically and provide a comprehensive report on the server's performance, any issues found, and recommendations for optimization. ``` --- ## Alternative Shorter Test Prompt If you want a quicker test, use this condensed version: ``` Test my Calibre Library MCP server (100k+ books): 1. Check connection and get library stats 2. Search for "Stephen King" books 3. Get batch details for 20 random books 4. Extract content from book ID 1 5. Try full-text search for "magic" across all books 6. Test edge case: invalid book ID 999999 For each test, report: response time, accuracy, any errors. Focus on performance with this large library. ``` --- ## Performance Benchmarks to Watch For When running these tests, look for: **✅ Good Performance:** - Simple searches: < 1 second - Batch operations (20-50 books): < 2 seconds - Content extraction: < 3 seconds - Pagination: < 1 second per page **⚠️ Warning Signs:** - Any operation > 5 seconds - Memory errors or timeouts - Inconsistent response formats - Missing pagination info **🚨 Critical Issues:** - Connection failures - Database errors - Crashes or hangs - Incorrect result counts --- ## Usage Instructions 1. **Configure Claude Desktop** with your MCP server: **For Windows (Recommended - using your virtual environment):** ```json { "mcpServers": { "calibre-library": { "command": "C:\\path\\to\\your\\project\\.venv\\Scripts\\python.exe", "args": ["calibre_mcp_server.py"], "cwd": "C:\\path\\to\\your\\project", "env": { "CALIBRE_CONFIG_PATH": "C:\\path\\to\\your\\project\\config.local.json", "PYTHONUTF8": "1" } } } } ``` **Alternative - using system Python (if venv doesn't work):** ```json { "mcpServers": { "calibre-library": { "command": "python", "args": ["calibre_mcp_server.py"], "cwd": "C:\\path\\to\\your\\project", "env": { "CALIBRE_CONFIG_PATH": "C:\\path\\to\\your\\project\\config.local.json", "PYTHONUTF8": "1" } } } } ``` **For Linux/Mac (if testing on other systems):** ```json { "mcpServers": { "calibre-library": { "command": "/path/to/your/project/.venv/bin/python", "args": ["calibre_mcp_server.py"], "cwd": "/path/to/your/project", "env": { "CALIBRE_CONFIG_PATH": "/path/to/your/project/config.local.json" } } } } ``` 2. **Copy the main test prompt** into a new Claude Desktop conversation 3. **Let Claude run through all tests systematically** 4. **Review the results** for performance issues 5. **Note any specific areas needing optimization** This will give you a comprehensive evaluation of your server's readiness for production use with large-scale libraries.