Migrating from FlexSearch to rapid-fuzzy

FlexSearch is a full-text search engine that uses inverted indexes and tokenization. rapid-fuzzy is a fuzzy matching library: its search matches items that contain the query’s characters in order (so tsc and typscript find TypeScript), and it provides edit-distance and similarity functions. They serve different primary use cases, but rapid-fuzzy can replace FlexSearch for autocomplete over short strings.

When to switch

Switch to rapid-fuzzy when:

• You need matches for partial words, abbreviations and missing letters (e.g., “tsc” or “typscript” → “TypeScript”)
• You want ranked results by string similarity
• You need match highlighting positions
• You search short strings (names, tags, file names, commands)

Stay with FlexSearch when:

• You search full-text documents (articles, blog posts, logs)
• You rely on tokenization, stemming, or language-specific analyzers
• You need field-level scoring with document weights

Installation

# Remove FlexSearch
npm uninstall flexsearch

# Install rapid-fuzzy
npm install rapid-fuzzy

API Mapping

FlexSearch	rapid-fuzzy	Notes
new Index() + index.add(id, text)	new FuzzyIndex(items)	rapid-fuzzy indexes by array position
index.search(query)	index.search(query)	Returns { item, score, index }[]
index.add(id, text)	index.add(item)	Appends to index
index.remove(id)	index.remove(index)	Removes by position; the last item moves into the freed position

Code Examples

Basic search

const items = ['TypeScript', 'JavaScript', 'Python'];

// Before (FlexSearch)
import { Index } from 'flexsearch';
const index = new Index();
items.forEach((item, i) => index.add(i, item));
index.search('typescript'); // [0] — returns IDs only
index.search('typscript');  // [] — not a token of any item

// After (rapid-fuzzy)
import { FuzzyIndex } from 'rapid-fuzzy';
const index = new FuzzyIndex(items);
index.search('typscript');
// [{ item: 'TypeScript', score: 0.86, index: 0, positions: [] }]

Standalone search (no index)

// Before (FlexSearch) — always requires index setup
import { Index } from 'flexsearch';
const index = new Index();
items.forEach((item, i) => index.add(i, item));
const ids = index.search(query);

// After (rapid-fuzzy) — one-liner for simple cases
import { search } from 'rapid-fuzzy';
const results = search(query, items);

Options

// Before (FlexSearch)
index.search(query, { limit: 10 });

// After (rapid-fuzzy)
index.search(query, { maxResults: 10, minScore: 0.3 });

What You Gain

Fuzzy matching

FlexSearch’s default index matches whole tokens — “typscript” won’t find “TypeScript”. rapid-fuzzy’s search finds items that contain the query’s characters in order, so partial words, abbreviations and missing letters match. It does not tolerate substituted or swapped letters (tpyescript); for those, compare candidates with a distance function such as jaroWinklerMany or damerauLevenshtein.

Similarity scores

Every result includes a normalized score (0.0–1.0), enabling threshold-based filtering and ranked display.

Match positions

Get character-level match positions for highlighting:

import { search } from 'rapid-fuzzy';
import { highlight } from 'rapid-fuzzy/highlight';

const results = search('type', items, { includePositions: true });
const html = highlight(results[0].item, results[0].positions, '<b>', '</b>');

Additional algorithms

Access edit-distance and similarity functions for specialized use cases:

import { levenshtein, jaroWinkler, sorensenDice } from 'rapid-fuzzy';