English
Skip to contentOne dictionary.
Aho–Corasick / exact text matching
One dictionary.
One scan.
Every match.
Match a dictionary against text with grapheme precision and slice-ready ranges. Follow every transition, fallback, and suffix match.
he · she · his · hers
ushers → she, he
ushers → hers
ushers → she, he
ushers → hers
search('ushers')
{ pattern: 'she', start: 1, end: 4 }
text.slice(1, 4) === 'she'
{ pattern: 'she', start: 1, end: 4 }
text.slice(1, 4) === 'she'
Documentation for v3.2.0, available on npm. MIT · SonOfMagic · BrunoRB/ahocorasick
3.2.0current release
grapheme-firstmatching unit
UTF-16slice-ready ranges
0runtime dependencies
Start with a task
A focused API for each kind of text work.
Pick the result you need, then keep the same matcher as your application grows.
Find
Search every occurrence
Get independent matches with original text ranges, duplicate pattern indexes, and overlap control.
matcher.search(text)MeasureCheck, count, aggregate
Stop on the first hit, count all occurrences, or count each dictionary entry without creating match objects.
matcher.countByPattern(text)RenderHighlight safely
Tokenize non-overlapping matches and render user input as text nodes, including strings that look like HTML.
matcher.tokenize(text)StreamProcess chunks
Carry graphemes across chunk boundaries with sessions, async iterables, previews, and backpressure.
createTokenStream(matcher)The contract to remember
Match the original text. Keep the original coordinates.
Every result uses a half-open UTF-16 range, so the matched text is always text.slice(start, end). Emoji, combining marks, and CRLF stay together as graphemes.
const text = '😀 café'
const hit = new AhoCorasick(['café']).search(text)[0]
text.slice(hit.start, hit.end)
// 'café'Keep exploring