Skip to content
Recombee Logo

Personalized

Semantic Search

That Understands What Users Mean

Start Free Request Demo

Proven Results

Measurable Growth Across Industries

Live Demo

Search by Meaning, Not Just Keywords

Try it -> Find a movie in your own words.

Live results from Recombee Media Demo.

Key Features

Search That Learns From Every User

Semantic Understanding

Understand the intent and context behind every query. An LLM helps users find what they’re looking for, even when their search terms don’t exactly match the content, such as “feel-good movies for a rainy evening” or “warm jackets for extreme cold”.

Personalized Results

Go beyond query relevance to reflect individual preferences. Machine learning uses each user’s interaction history and interests to personalize search rankings, so two people searching for the same thing see results tailored to them.

Real-Time Adaptability

Keep search results fresh within milliseconds. React to new user interactions and catalog changes in real time, without waiting for scheduled model retraining.

Full Control & Flexibility

Configure every search box as a Scenario directly in the Admin UI, without code changes. Choose the search Logic, apply ReQL filters and boosters to align results with business goals, and let the AI ReQL Assistant turn plain-language rules into working queries.

Scalability

Handle massive search volumes without compromising speed. Horizontally scalable infrastructure supports over 70k recommendations per second and catalogs of tens of millions of items, with globally distributed data centers ensuring low latency.

The Building Blocks of Your Search

Semantic Search

Understand the intent and context behind every query. An LLM helps users find what they’re looking for, even when their search terms don’t exactly match the content.

Quick Search

Help users find what they’re looking for faster with suggestions as they type. Surface relevant items based on their query so far, personalized to each user’s interests.

Faceted Search

Make it easy to narrow down results by attributes such as brand, size, price, or category, while keeping results personalized to each user.

Search Item Segments

Go beyond individual items and help users discover entire groups of related content or products, from categories and brands to genres and artists.

Personalized Ranking

Put the most relevant results first for every user. Use individual interaction history and preferences to personalize rankings, so the same query can deliver different results to different people.

Search Synonyms

Help users find what they’re looking for, whatever terms they use. Define synonyms in the Admin UI or through the API to connect different search terms to the same results.

A/B Testing

Find out which search strategies work best before rolling them out to everyone. Test different search variants and configurations to compare their impact on click-through rates, conversions, and other key metrics.

ReQL

Control exactly what appears in search results with flexible filters and boosters. Define business rules, from hiding out-of-stock items to promoting new arrivals, and let the AI ReQL Assistant turn plain English into working queries.

A Few Steps
to Personalization

11SEND DATACatalogItem CatalogUser CatalogInteractionsLive InteractionsInteraction HistoryREAL-TIME22SET UP SCENARIOSSCENARIORecommendation TypeRecommendation LogicFilters & BoostersConstraints33DELIVER PERSONALIZATIONSemantic SearchQuick SearchFaceted Search44MEASURE & OPTIMIZEANALYTICS & INSIGHTSReal-Time KPIsInsightsA/B Testing

Ship Personalized Search With or Without Code

AI Agents Ready

If you’re using an AI agent or assistant to integrate Recombee into your application, the documentation includes an integration guide written specifically for AI assistants.

Use the ready-made prompt below as a starting point.

Initial Prompt

Integrate Recombee (a recommendation engine) into my application.

Before writing any code, fetch and read the integration guide at https://docs.recombee.com/ai_assistant_guide.md.
It's written specifically for AI assistants and covers authentication, catalog sync, interactions, recommendations, and common pitfalls.
Use it as your primary reference.

First, explore the codebase to understand the stack, data models, and integration points.
Then propose a plan before implementing anything.

Do not decide on your own where recommendations should appear.
Ask me which use cases to implement — homepage recommendations, related items on a product page, personalized search, email suggestions, and so on.
The right surfaces depend on my product and priorities, so treat that as a shared decision even when you're otherwise working autonomously.

After we agree on the use cases, implement them end to end, and flag any assumptions you have to make along the way.

Personalization Powerhouse works best as a whole.

Real-Time
Recommendations
Real-Time
Analytics & Insights

Shape the Future of Personalization

Start Your Free 30-Day Trial Today.

Start Free
  • Improve Key Metrics
  • Drive User Engagement
  • Maximize Profits
  • Boost Customer Loyalty