/**
 * URL Metadata Fetching
 *
 * Scrape URLs to extract metadata: title, description, OpenGraph images,
 * JSON-LD structured data, creator info, categories, etc.
 * Results are cached to avoid re-fetching.
 *
 * Features:
 * - Parallel scraping (default 5 concurrent)
 * - Hard timeout per URL (default 4s)
 * - Caching via ResponseCache interface
 */
import type { ResponseCache } from '../caching/types';
import type { CandidateMessage } from '../types';
import type { ScrapedMetadata, ScraperConfig } from './types';
/**
 * Extract all unique URLs from text.
 */
export declare function extractUrlsFromText(text: string): string[];
/**
 * Extract all unique URLs from an array of candidates.
 * Looks in the candidate content and surrounding context.
 */
export declare function extractUrlsFromCandidates(candidates: readonly CandidateMessage[]): string[];
interface FetchOptions extends ScraperConfig {
    /** Callback when scraping starts (only called if there are uncached URLs) */
    onScrapeStart?: ((info: {
        urlCount: number;
        cachedCount: number;
    }) => void) | undefined;
    /** Callback for each URL scraped (always called - success or failure) */
    onUrlScraped?: ((info: {
        url: string;
        success: boolean;
        error?: string | undefined;
        current: number;
        total: number;
    }) => void) | undefined;
    /** Callback for debug logging */
    onDebug?: ((message: string) => void) | undefined;
    /** Max concurrent scrapes (default 5) */
    concurrency?: number | undefined;
    /** Cache for scrape results */
    cache?: ResponseCache | undefined;
}
/**
 * Fetch metadata for a list of URLs.
 * Best-effort: failures are silently skipped.
 * Uses parallel scraping with configurable concurrency.
 * Caches both successes and failures to avoid re-fetching.
 */
export declare function fetchMetadataForUrls(urls: readonly string[], options?: FetchOptions): Promise<Map<string, ScrapedMetadata>>;
export {};
//# sourceMappingURL=metadata.d.ts.map