/**
 * i18n word tokenizer for the L1 lexical index. Uses Intl.Segmenter
 * (standard ECMAScript — hub-portable) to dictionary-segment Thai/Lao/Khmer/CJK,
 * which the word-run `tokenize` cannot. Terms are matched in NFKC-lowercased form;
 * offsets index the ORIGINAL text so snippets slice the user's text.
 */
import type { Tokenizer } from './tokenize.js';
export interface Token {
    readonly term: string;
    readonly offset: number;
}
/** Word-like tokens with NFKC-lowercased terms + original-text char offsets. */
export declare const segmentTokens: (text: string) => Token[];
/** Term-only tokenizer (the public `Tokenizer` shape) — for queries. */
export declare const segmentTokenizer: Tokenizer;