Doc Agent

#!/usr/bin/env node import { resolve } from 'node:path'; import chalk from 'chalk'; import { Command } from 'commander'; import { runExtract } from './commands/extract'; // Resolve paths relative to where user ran the command // INIT_CWD is set by pnpm to original working directory const cwd = process.env.INIT_CWD || process.cwd(); function resolvePath(filePath: string): string { return resolve(cwd, filePath); } const program = new Command(); program .name('doc') .alias('doc-agent') .description('Document extraction and semantic search CLI') .version('0.1.0'); program .command('extract <file>') .description('Extract structured data from a document') .option('-p, --provider <provider>', 'AI provider (gemini|openai|ollama)', 'ollama') .option('-m, --model <model>', 'Model to use (ollama: llama3.2-vision, gemini: gemini-2.5-flash)') .option('-d, --dry-run', 'Print JSON only, do not save to database', false) .action(async (file: string, options) => { const absolutePath = resolvePath(file); // Set default model based on provider if not specified const defaultModels: Record<string, string> = { ollama: 'llama3.2-vision', gemini: 'gemini-2.5-flash', openai: 'gpt-4o', }; const model = options.model || defaultModels[options.provider] || 'llama3.2-vision'; await runExtract(absolutePath, { provider: options.provider, model, dryRun: options.dryRun, }); }); program .command('search <query>') .description('Search indexed documents') .action(async (query: string) => { console.log(chalk.yellow('Search not yet implemented')); console.log('Query:', query); }); program .command('index <directory>') .description('Index all documents in a directory') .action(async (directory: string) => { console.log(chalk.yellow('Batch indexing not yet implemented')); console.log('Directory:', directory); }); program .command('mcp') .description('Start MCP server') .action(async () => { const { startMCPServer } = await import('./mcp/server.js'); await startMCPServer(); }); program.parse();