Hybrid Search with Manticore Search

Combine full-text and vector retrieval in Manticore Search for more relevant results.

What is Hybrid Search?

Hybrid search combines full-text and vector search in a single query. It helps you handle both exact keywords and semantic meaning at the same time, so results can match identifiers, product names, and error codes while still understanding natural-language intent.

When to use Hybrid Search?

Improving search relevance for natural-language queries
Building RAG pipelines that need stronger retrieval
Combining exact keyword matching with semantic similarity
Searching product catalogs with names, SKUs, and descriptions
Searching support knowledge bases with error codes and symptoms
Handling queries that mix identifiers with descriptive text
Improving content discovery in large document collections
Ranking search results using both lexical precision and semantic recall
Searching multilingual or synonym-rich content
Implementing AI-powered search without losing exact-match behavior

Why Manticore Search is good for Hybrid Search

Manticore Search supports hybrid search natively, combining MATCH() and KNN() in one query.
It works through both SQL and JSON interfaces, making integration straightforward.
You can keep exact-match precision for identifiers while adding semantic understanding for natural language.
Multiple KNN subqueries can be used in the same hybrid query when needed.
Manticore lets you build hybrid retrieval without adding a separate search engine.

How to get started

Install Manticore Search

Visit the official Manticore Search website: https://manticoresearch.com/
Follow the installation instructions for your operating system
Alternatively, use Docker: docker pull manticoresearch/manticore

Prepare your data

Create a table with text fields for full-text search
Add vector fields or configure auto-embeddings for semantic retrieval
Index your documents so both text and vector signals are available

Run your first hybrid query

Use MATCH() together with KNN() in SQL, or the equivalent JSON request
Test queries that combine exact terms with natural-language intent
Review the fused results and confirm they improve relevance

Tune your hybrid search

Adjust your full-text query and vector query inputs
Apply attribute filters to keep results within the right category or tenant
Experiment with multiple KNN subqueries if your data has several semantic dimensions

Enjoy Hybrid Search

Use Manticore Search to deliver more relevant search results
Feel free to create an issue if you encounter any problems
Consider our professional services for advanced implementations

Pros

✓ Combines exact matching with semantic understanding

✓ Improves relevance for real-world queries that mix keywords and intent

✓ Supports both SQL and JSON interfaces

✓ Works with vector fields and auto-embeddings

✓ Can use multiple KNN subqueries in one hybrid query

✓ Reduces the need for custom query orchestration

Cons

✗ Requires vector data or auto-embeddings in addition to full-text fields

✗ May need tuning to balance lexical and semantic signals

✗ Consumes more resources than plain full-text search alone

Resources

Learn more about other use cases

Explore the other Manticore Search use cases to see how one engine handles full-text, vector, hybrid, autocomplete, and log-search workloads.

Install Manticore Search for Hybrid Search

Try Manticore Search for implementing Hybrid Search in your applications today!

Install Now