Presearch MCP Server

import { createMcpServer } from './src/mcp-server.js'; import { loadConfig } from './src/core/config.js'; async function demonstrateAiAgentAccess() { console.log('🤖 Demonstrating How AI Agents Access MCP Content...\n'); const config = loadConfig(); try { // Create MCP server instance (this is what AI clients connect to) const server = await createMcpServer(config); console.log('✅ MCP Server Created Successfully'); console.log('📋 Server Info:'); console.log(` - Name: ${server.name}`); console.log(` - Version: ${server.version}`); console.log(` - Description: ${server.description}`); // Simulate what an AI agent would do console.log('\n🔍 Simulating AI Agent Request...'); // Example 1: Simple search (what Claude/Cursor would call) console.log('\n1️⃣ AI Agent: "Search for latest AI news"'); const searchTool = server._tools.find(t => t.name === 'presearch_ai_search'); if (searchTool) { console.log(' ✅ Found search tool'); console.log(' 📝 Tool Schema:', JSON.stringify(searchTool.inputSchema, null, 4)); // Execute search (simulating AI agent call) const searchResults = await searchTool.execute({ query: "latest AI breakthroughs December 2024", limit: 3 }); console.log(' 📊 Results received:', searchResults.results.length, 'items'); console.log(' 📄 First result:', { title: searchResults.results[0].title, url: searchResults.results[0].url, snippet: searchResults.results[0].snippet.substring(0, 100) + '...' }); } // Example 2: Export functionality console.log('\n2️⃣ AI Agent: "Export these results to JSON"'); const exportTool = server._tools.find(t => t.name === 'export_search_results'); if (exportTool) { console.log(' ✅ Found export tool'); // Create sample results for export const sampleResults = [ { title: "Sample AI Article", url: "https://example.com/ai-article", snippet: "This is a sample article about AI breakthroughs...", description: "Detailed article content...", source: "example.com", publishedDate: "2024-12-14" } ]; const exportResult = await exportTool.execute({ results: sampleResults, format: "json", filename: "ai_agent_export_test.json" }); console.log(' 💾 Export result:', exportResult.content); console.log(' 📁 File saved to:', process.cwd() + '/ai_agent_export_test.json'); } // Example 3: Deep research (complex AI agent workflow) console.log('\n3️⃣ AI Agent: "Conduct deep research on quantum computing"'); const researchTool = server._tools.find(t => t.name === 'presearch_deep_research'); if (researchTool) { console.log(' ✅ Found deep research tool'); console.log(' 🎯 Starting research (this may take a moment)...'); const researchResults = await researchTool.execute({ query: "quantum computing applications 2024", depth: 1, breadth: 2, research_focus: "technology" }); console.log(' 📚 Research completed:', researchResults.success); if (researchResults.success) { console.log(' 📖 Research summary:', researchResults.research_summary.substring(0, 200) + '...'); console.log(' 🔗 Sources analyzed:', researchResults.sources_analyzed); } } console.log('\n✅ AI Agent Access Demonstration Complete!'); console.log('\n📋 Summary:'); console.log(' • AI agents connect to MCP server via stdio or HTTP'); console.log(' • Server exposes tools that AI can call with parameters'); console.log(' • Results are returned as JSON for AI to process'); console.log(' • Files are saved locally when export tools are used'); console.log(' • AI agents get structured data, not raw HTML'); } catch (error) { console.error('❌ Demonstration failed:', error.message); if (error.response) { console.error('API Response:', error.response.data); } } } // Run the demonstration demonstrateAiAgentAccess().catch(err => { console.error('Fatal Error:', err); process.exit(1); });