@just-every/search/dist/cli.js

Search every provider at once - Google, Anthropic, OpenAI, Perplexity, X.AI and Brave.

#!/usr/bin/env node
import 'dotenv/config';
import yargs from 'yargs';
import { hideBin } from 'yargs/helpers';
import { web_search, web_search_task } from './index.js';
function isImageQuery(query) {
    const imageKeywords = [
        'image', 'images', 'photo', 'photos', 'picture', 'pictures', 'pic', 'pics',
        'logo', 'logos', 'icon', 'icons', 'screenshot', 'screenshots',
        'jpg', 'jpeg', 'png', 'gif', 'svg', 'webp', 'bmp',
        'wallpaper', 'wallpapers', 'background', 'backgrounds',
        'thumbnail', 'thumbnails', 'avatar', 'avatars'
    ];
    const lowerQuery = query.toLowerCase();
    return imageKeywords.some(keyword => lowerQuery.includes(keyword));
}
const argv = yargs(hideBin(process.argv))
    .command('$0 <query>', 'Search the web using various search engines', (yargs) => {
    return yargs
        .positional('query', {
        describe: 'The search query',
        type: 'string',
        demandOption: true,
    })
        .option('engine', {
        alias: 'e',
        describe: 'Search engine to use',
        type: 'string',
        choices: ['brave', 'brave-images', 'anthropic', 'openai', 'google', 'sonar', 'sonar-pro', 'sonar-deep-research', 'xai'],
        default: 'brave',
    })
        .option('results', {
        alias: 'n',
        describe: 'Number of results to return',
        type: 'number',
        default: 5,
    })
        .option('json', {
        alias: 'j',
        describe: 'Output raw JSON',
        type: 'boolean',
        default: false,
    });
}, async (argv) => {
    try {
        let engine = argv.engine;
        // Auto-detect image queries and suggest brave-images if using brave
        if (engine === 'brave' && isImageQuery(argv.query)) {
            console.log(`🔍 Detected image query. Consider using --engine brave-images for better results.\n`);
        }
        // CLI doesn't use inject_agent_id, so we use the backward-compatible signature
        const result = await web_search(engine, argv.query, argv.results);
        if (argv.json) {
            console.log(result);
        }
        else {
            if (result.startsWith('Error:')) {
                console.error(result);
                process.exit(1);
            }
            try {
                const results = JSON.parse(result);
                console.log(`\nSearch results for: "${argv.query}"\n`);
                results.forEach((item, index) => {
                    console.log(`${index + 1}. ${item.title}`);
                    console.log(`   ${item.url}`);
                    // Handle image results
                    if (item.thumbnail && item.source) {
                        console.log(`   Source: ${item.source}`);
                        console.log(`   Thumbnail: ${item.thumbnail}`);
                        if (item.width && item.height) {
                            console.log(`   Dimensions: ${item.width}x${item.height}`);
                        }
                    }
                    else if (item.snippet) {
                        console.log(`   ${item.snippet}`);
                    }
                    console.log();
                });
            }
            catch (e) {
                console.log(result);
            }
        }
    }
    catch (error) {
        console.error('Search failed:', error);
        process.exit(1);
    }
})
    .command('task <query>', 'Run comprehensive research using web_search_task', (yargs) => {
    return yargs
        .positional('query', {
        describe: 'The research query',
        type: 'string',
        demandOption: true,
    })
        .option('model-class', {
        alias: 'm',
        describe: 'Model class to use',
        type: 'string',
        choices: ['standard', 'mini', 'reasoning', 'reasoning_mini', 'monologue', 'metacognition', 'code', 'writing', 'summary', 'vision', 'vision_mini', 'image_generation', 'embedding', 'voice'],
        default: 'reasoning_mini',
    });
}, async (argv) => {
    try {
        console.log(`Running comprehensive research on: "${argv.query}"\n`);
        console.log(`Using model class: ${argv.modelClass}\n`);
        const result = await web_search_task(argv.query, argv.modelClass);
        if (result.startsWith('Error:')) {
            console.error(result);
            process.exit(1);
        }
        console.log(result);
    }
    catch (error) {
        console.error('Research task failed:', error);
        process.exit(1);
    }
})
    .help()
    .parse();
//# sourceMappingURL=cli.js.map