Search that understands what people mean
Users type in their own words and still find the right result, in 100+ languages.
100 free credits on sign up. No card required.
Superman: The Complete Story
A boy from another planet grows up with the power of flight.
The Rocketeer
A stunt pilot finds a jet pack and takes to the skies.
Peter Pan
A boy who never grows up flies children to Neverland.
Different words, same meaning. “Affordable running shoes” finds budget sneakers and cheap trainers.
Budget sneakers for daily runs
Cheap trainers under $60
Best value road running shoes
Typos still work. Small spelling mistakes don't end in empty results.
wireles hedphones
Wireless Headphones, Noise Cancelling
Any language in, any language out. A Korean query finds an English document.
가성비 노트북
Best budget laptops of the year
Long, chatty queries work. So do short, vague ones.
Product codes and names still match. Keyword search runs alongside.
One query, both kinds of matches. Meaning and exact terms ranked together.
Query: “iPhone 15 Pro case”
iPhone 15 Pro Silicone Case with MagSafe
Protective cover for iPhone 15 Pro Max
Clear MagSafe Case, iPhone 15 Pro
Keyword search matches words like “affordable” and “laptop”. Vecstore matches what the person actually wants.
Keyword search
Affordable Laptop Stand
School Laptop Policy
Laptop Sleeve, Affordable Neoprene
Vecstore
Best Budget Laptops for Students
Cheap Notebooks for College
Student Laptop Buying Guide
cant log in
Reset your password
shoes for flat feet
Stability running shoes
who approves expenses
Expense policy, 2025
Search across languages. A query in one finds documents in another.
No setup per language. Same API, same index, nothing to configure.
100+
Supported languages include
It finds results by meaning instead of exact words, so "cheap running shoes" matches "budget sneakers". You send a query and get ranked results back.
Any text you add: product catalogs, help docs, articles, support tickets or user posts.
No. Vecstore runs the models and the vector database for you. You call one API.
In our tests on 10 million records, a search took 3-4ms. Turning the query into a vector adds about 20ms.

3–4ms average search latency
100 Free credits. No credit card required.