MCP Deep Search

#!/usr/bin/env node // Load environment variables BEFORE importing other modules import * as dotenv from 'dotenv'; dotenv.config(); import { Command } from 'commander'; import { web_search, web_search_task } from '@just-every/search'; import { readFileSync, writeFileSync } from 'fs'; import { fileURLToPath } from 'url'; import { dirname, join } from 'path'; const __filename = fileURLToPath(import.meta.url); const __dirname = dirname(__filename); // Read package.json for version const packageJson = JSON.parse( readFileSync(join(__dirname, '../package.json'), 'utf-8') ); const program = new Command(); program .name('mcp-deep-search') .description('MCP server for deep web search using @just-every/search') .version(packageJson.version); program .command('search <query>') .description('Perform a web search') .option( '-p, --provider <provider>', 'Search provider (brave, anthropic, openai, google, etc.)', 'brave' ) .option('-n, --max-results <number>', 'Maximum number of results', '10') .option('-a, --include-answer', 'Include AI-generated answer if available') .option('-o, --output <path>', 'Output results to file (JSON format)') .action(async (query: string, options) => { try { console.error(`🔍 Searching for: "${query}"`); console.error(`📍 Provider: ${options.provider}`); const resultsJson = await web_search( options.provider || 'brave', query, parseInt(options.maxResults, 10) ); let results; // Check if the response is an error if (resultsJson.startsWith('Error:')) { throw new Error(resultsJson); } // Try to parse as JSON try { results = JSON.parse(resultsJson); // Handle the new format where results are directly an array if (Array.isArray(results)) { results = { results: results }; } } catch { // If not JSON, treat as plain text answer results = { answer: resultsJson, results: [], }; } // Format output const output = { query, provider: options.provider, timestamp: new Date().toISOString(), answer: results.answer, results: results.results, }; if (options.output) { // Save to file writeFileSync(options.output, JSON.stringify(output, null, 2)); console.error(`✅ Results saved to: ${options.output}`); } else { // Output to console if (results.answer) { console.log('\n📝 AI Answer:'); console.log(results.answer); console.log('\n---\n'); } if (results.results && results.results.length > 0) { console.log('🔍 Search Results:\n'); results.results.forEach((result: any, index: number) => { console.log(`${index + 1}. ${result.title}`); console.log(` URL: ${result.url || result.link}`); if (result.snippet) { console.log(` ${result.snippet}`); } console.log(); }); } console.error( `\n✅ Found ${results.results?.length || 0} results` ); } } catch (error) { console.error( '❌ Error:', error instanceof Error ? error.message : error ); process.exit(1); } }); program .command('research <query>') .description( 'Perform comprehensive research using multiple search engines automatically' ) .option( '-m, --model-class <class>', 'AI model class to use', 'reasoning_mini' ) .option('-o, --output <path>', 'Output report to file') .action(async (query: string, options) => { try { console.error(`🔬 Starting comprehensive research on: "${query}"`); console.error(`🤖 Using model class: ${options.modelClass}\n`); const report = await web_search_task(query, options.modelClass); if (options.output) { // Save to file writeFileSync(options.output, report); console.error( `\n✅ Research report saved to: ${options.output}` ); } else { // Output to console console.log(report); } } catch (error) { console.error( '❌ Error:', error instanceof Error ? error.message : error ); process.exit(1); } }); program .command('serve') .description('Run as an MCP server') .action(async () => { // Import and run the serve module await import('./serve.js'); }); program.parse();