code example

Build Multilingual Search with altor-vec

What this pattern solves: Cross-language retrieval when users search in one language and the corpus lives in another.

Global products need search that survives translation gaps. A Spanish query such as 'restablecer contraseña' should still find English support material if the underlying embedding space aligns those meanings.

Local retrieval helps international apps keep the same fast UX everywhere, even where mobile connectivity is inconsistent or where the corpus is distributed as a static bundle.

Install

npm install altor-vec

Concept explanation

In a multilingual search workflow, users usually describe intent in their own words. That is why vector search works well here: each record is turned into an embedding, the embeddings are indexed once, and later queries retrieve the nearest semantic neighbors instead of relying only on exact tokens. In practice this means the interface can respond to paraphrases, shorthand, and partial descriptions far better than a literal-only search box.

The browser is often the right place to do this when the corpus is moderate in size and safe to ship. The instant benefit is lower latency. The architectural benefit is that you remove a whole search service from the request path. That matters for keystroke-heavy interactions, offline-capable apps, and product surfaces where search should feel like a UI primitive rather than a network round trip.

This page uses a deterministic embedding helper so the sample is runnable with only altor-vec installed. That keeps the example honest and easy to paste into a demo project. The key is the embedding model, not the ANN library. altor-vec handles the nearest-neighbor lookup after you choose a multilingual embedding model or precompute aligned vectors.

Representative browser benchmark: ~54KB gzipped library payload, sub-millisecond local query time on a moderate corpus, and no per-query API dependency. Exact numbers depend on vector dimensions, index parameters, and device class.

Runnable JavaScript example

The following snippet indexes a small in-memory dataset, performs a semantic lookup for restablecer contraseña del equipo, and prints the nearest matches. It uses the real altor-vec API, including init(), WasmSearchEngine.from_vectors(), and search().

import init, { WasmSearchEngine } from 'altor-vec';

        const dims = 12;
        const records = [
  {
    "title": "Reset workspace password",
    "text": "Instructions for changing or resetting a forgotten account password.",
    "meta": "account"
  },
  {
    "title": "Invite a teammate",
    "text": "How to add another person and assign the right access role.",
    "meta": "team"
  },
  {
    "title": "Download invoices",
    "text": "Where to find paid invoices and billing history exports.",
    "meta": "billing"
  },
  {
    "title": "Actualizar método de pago",
    "text": "Pasos para cambiar la tarjeta asociada a una suscripción activa.",
    "meta": "billing"
  },
  {
    "title": "Configurer l'authentification SSO",
    "text": "Guide d'activation du SAML pour les connexions d'entreprise.",
    "meta": "security"
  },
  {
    "title": "Esportare i dati",
    "text": "Come creare esportazioni CSV o JSON dal pannello dati.",
    "meta": "data"
  }
];

        function embedText(text) {
  const vector = new Float32Array(dims);
  for (const token of text.toLowerCase().split(/[^a-z0-9]+/).filter(Boolean)) {
    let hash = 2166136261;
    for (const char of token) {
      hash = Math.imul(hash ^ char.charCodeAt(0), 16777619);
    }
    const slot = Math.abs(hash) % dims;
    vector[slot] += 1;
    vector[(slot + token.length) % dims] += token.length / 10;
  }
  const magnitude = Math.hypot(...vector) || 1;
  return Array.from(vector, (value) => value / magnitude);
}

        async function main() {
          await init();

          const flat = new Float32Array(
            records.flatMap((record) => embedText(`${record.title} ${record.text} ${record.meta}`))
          );

          const engine = WasmSearchEngine.from_vectors(flat, dims, 16, 200, 64);
          const hits = JSON.parse(engine.search(new Float32Array(embedText('restablecer contraseña del equipo')), 4));

          const results = hits.map(([id, distance]) => ({
            ...records[id],
            similarity: Number((1 - distance).toFixed(3)),
          }));

          console.table(results);
          engine.free();
        }

        main();

React component version

The React version keeps the same index build but wires it into component state so the UI can query on input changes. That is usually how teams introduce semantic retrieval into an existing product: initialize once, keep the engine in memory, and map nearest-neighbor hits back to the original records.

import { useEffect, useState } from 'react';
        import init, { WasmSearchEngine } from 'altor-vec';

        const dims = 12;
        const records = [
  {
    "title": "Reset workspace password",
    "text": "Instructions for changing or resetting a forgotten account password.",
    "meta": "account"
  },
  {
    "title": "Invite a teammate",
    "text": "How to add another person and assign the right access role.",
    "meta": "team"
  },
  {
    "title": "Download invoices",
    "text": "Where to find paid invoices and billing history exports.",
    "meta": "billing"
  },
  {
    "title": "Actualizar método de pago",
    "text": "Pasos para cambiar la tarjeta asociada a una suscripción activa.",
    "meta": "billing"
  },
  {
    "title": "Configurer l'authentification SSO",
    "text": "Guide d'activation du SAML pour les connexions d'entreprise.",
    "meta": "security"
  },
  {
    "title": "Esportare i dati",
    "text": "Come creare esportazioni CSV o JSON dal pannello dati.",
    "meta": "data"
  }
];

        function embedText(text) {
  const vector = new Float32Array(dims);
  for (const token of text.toLowerCase().split(/[^a-z0-9]+/).filter(Boolean)) {
    let hash = 2166136261;
    for (const char of token) {
      hash = Math.imul(hash ^ char.charCodeAt(0), 16777619);
    }
    const slot = Math.abs(hash) % dims;
    vector[slot] += 1;
    vector[(slot + token.length) % dims] += token.length / 10;
  }
  const magnitude = Math.hypot(...vector) || 1;
  return Array.from(vector, (value) => value / magnitude);
}

        export function MultilingualSearchExample() {
          const [engine, setEngine] = useState(null);
          const [query, setQuery] = useState('');
          const [results, setResults] = useState([]);

          useEffect(() => {
            let cancelled = false;
            let instance;

            (async () => {
              await init();
              const flat = new Float32Array(
                records.flatMap((record) => embedText(`${record.title} ${record.text} ${record.meta}`))
              );
              instance = WasmSearchEngine.from_vectors(flat, dims, 16, 200, 64);
              if (!cancelled) setEngine(instance);
            })();

            return () => {
              cancelled = true;
              instance?.free();
            };
          }, []);

          useEffect(() => {
            if (!engine || query.trim().length < 2) {
              setResults([]);
              return;
            }

            const hits = JSON.parse(engine.search(new Float32Array(embedText(query)), 5));
            setResults(
              hits.map(([id, distance]) => ({
                ...records[id],
                similarity: Number((1 - distance).toFixed(3)),
              }))
            );
          }, [engine, query]);

          return (
            <section>
              <input
                value={query}
                onChange={(event) => setQuery(event.target.value)}
                placeholder="Search in any supported language"
              />
              <ul>
                {results.map((result) => (
                  <li key={result.title}>
                    <strong>{result.title}</strong> — {result.meta} (score {result.similarity})
                  </li>
                ))}
              </ul>
            </section>
          );
        }

How this example works

The pattern has three moving parts. First, you choose what text represents each record: title, description, metadata, or a chunk of content. Second, you turn that text into vectors and flatten them into one Float32Array. Third, you build the HNSW graph and query it with a vector created from the user input. The library returns nearest-neighbor IDs and distances, and your app decides how to display or post-process them.

Because the retrieval step is approximate nearest-neighbor search, it stays fast even as the dataset grows beyond trivial linear scans. The most important quality lever is still the embedding model. Better vectors usually matter more than micro-optimizing ANN parameters, so teams should benchmark retrieval quality on real user phrasing before shipping the experience widely.

When to use this pattern

This is a practical fit when the search corpus is small to medium, shipped with the app, and searched frequently enough that backend latency would be noticeable. Common examples include docs portals, embedded support widgets, local-first assistants, and curated catalogs.

Global help centers
Travel apps
Consumer apps with multilingual UI
International internal tools

Limitations

The demo helper is not actually multilingual, so it only illustrates the retrieval wiring. Real multilingual quality depends on aligned embeddings and often benefits from evaluation across your supported languages.

Be especially careful about corpus size, update frequency, and data sensitivity. Browser vector search is excellent when those three constraints are favorable, but it is not the right answer when the dataset is huge, private, or changing constantly for every user.

FAQ

Is altor-vec itself multilingual?

The library is embedding-model agnostic. Multilingual behavior comes from the vectors you feed into it, not from the ANN engine alone.

Can one local index support many languages?

Yes, if the embeddings place semantically similar text from different languages near each other.

What should I benchmark first?

Recall quality across languages. Latency is usually easy; the harder part is validating that your embedding model really aligns the languages you care about.

Get started: npm install altor-vec · GitHub