code example

Build Semantic Filtering with altor-vec

What this pattern solves: Intent-based narrowing of results before or alongside structured filters.

Structured filters only work when the user understands your taxonomy. Semantic filtering helps them express fuzzy intent such as 'quiet ergonomic setup' or 'starter-friendly trail gear' before exact facets take over.

A local vector layer can score every item against a concept instantly, then your application can intersect that shortlist with normal filters like brand, stock, or price.

Install

npm install altor-vec

Concept explanation

In a semantic filtering 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. Think of semantic filtering as a complement to metadata filters rather than a replacement. The strongest experiences let both collaborate.

Runnable JavaScript example

The following snippet indexes a small in-memory dataset, performs a semantic lookup for beginner friendly camping gear, 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": "Starter campsite bundle",
    "text": "Beginner-friendly tent, sleeping bag, and lantern kit for weekend trips.",
    "meta": "camping"
  },
  {
    "title": "Ultralight thru-hike set",
    "text": "Minimal gear for experienced hikers prioritizing low pack weight.",
    "meta": "camping"
  },
  {
    "title": "Family car-camping tent",
    "text": "Large easy-pitch shelter with room dividers and storage pockets.",
    "meta": "camping"
  },
  {
    "title": "Portable camp stove",
    "text": "Compact gas stove for quick meals on short outdoor trips.",
    "meta": "cooking"
  },
  {
    "title": "Insulated sleeping pad",
    "text": "Warm inflatable pad for shoulder-season camping nights.",
    "meta": "sleep"
  },
  {
    "title": "Trail first-aid pouch",
    "text": "Compact medical kit for common cuts, blisters, and burns.",
    "meta": "safety"
  }
];

        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('beginner friendly camping gear')), 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": "Starter campsite bundle",
    "text": "Beginner-friendly tent, sleeping bag, and lantern kit for weekend trips.",
    "meta": "camping"
  },
  {
    "title": "Ultralight thru-hike set",
    "text": "Minimal gear for experienced hikers prioritizing low pack weight.",
    "meta": "camping"
  },
  {
    "title": "Family car-camping tent",
    "text": "Large easy-pitch shelter with room dividers and storage pockets.",
    "meta": "camping"
  },
  {
    "title": "Portable camp stove",
    "text": "Compact gas stove for quick meals on short outdoor trips.",
    "meta": "cooking"
  },
  {
    "title": "Insulated sleeping pad",
    "text": "Warm inflatable pad for shoulder-season camping nights.",
    "meta": "sleep"
  },
  {
    "title": "Trail first-aid pouch",
    "text": "Compact medical kit for common cuts, blisters, and burns.",
    "meta": "safety"
  }
];

        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 SemanticFilteringExample() {
          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="Describe the vibe or intent"
              />
              <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.

• Storefront refinement
• Marketplace discovery
• Curated collections
• Media library exploration

Limitations

Semantic scores are approximate and subjective. If users need exact eligibility rules, safety constraints, or guaranteed compliance filters, metadata must still remain authoritative.

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

What is semantic filtering good at?

It is good at translating fuzzy intent into a shortlist that structured filters can refine further.

Should semantic similarity override hard filters?

No. Use it to rank or narrow, but keep hard filters authoritative for rules like availability, compliance, or safety.

Can I expose the similarity score in the UI?

Sometimes, but many teams use it only behind the scenes and present the effect as 'smart matching' rather than a numeric score.

Get started: npm install altor-vec · GitHub