AI Embeddings · Vector Search

Your users can't find
what they're looking for.
AI search fixes that.

LIKE queries only match exact words. Users search in their own language — they use synonyms, context, intent. CodeNx embeds every record automatically, so search understands meaning, not just text.

Enable AI search → See live demo
⌕ handcrafted jewellery under ₹2000 ⟁ AI
⟁ Semantic results — 38ms 4 matches
Artisan Silver Bangles
Jewelry
97.2%
similarity
Blue Indigo Kurta (Light)
Apparel
94.8%
similarity
Sparkinity Soy Candle Set
Home & Living
91.5%
similarity
Zapvi Tempered Glass Pack
Accessories
89.3%
similarity
Old keyword search result
0 results for "handcrafted jewellery"
No exact match found. Try different keywords.

How it works

Three steps from dumb LIKE queries to intelligent semantic search.

1
◫
Define your collection
Add your fields in the schema builder. Select which text fields should contribute to the AI embedding — typically name, description, tags.
2
⟁
AI indexes automatically
Every record is vectorized in the background using state-of-the-art embeddings. New records embed on create. Edits re-index in seconds.
3
⚡
Query in plain language
POST to /api/{collection}/search with a natural language query. Get back ranked results with similarity scores. No DSL, no extra config.

Everything included in AI search

⟁
Multilingual support
Search queries and records in 100+ languages. Cross-language matching works out of the box — English query, Hindi records, no problem.
⧉
Filtered semantic search
Combine vector similarity with hard filters. Find 'products similar to X' within a specific category, price range, or status.
◈
Similarity recommendations
Expose a /similar endpoint to power 'you might also like' features. Just pass a record ID and get ranked similar records.
⎔
Auto re-indexing
Update a record's fields and its embedding updates automatically in seconds. Your search index is always fresh, zero manual effort.

Common questions about AI search

Give your users search
that actually works

Enable AI indexing on any collection and semantic search is live in minutes. No vector DB setup, no embedding infrastructure, no maintenance.

Start free — AI search included →

No credit card · Semantic search on free plan (up to 10K records)