What's the difference between lexical and semantic search?

Lexical search matches exact words and characters – if you search for "car", you only find documents containing the word "car". Semantic search understands meaning: searching for "car" also finds "vehicle", "automobile", and "sedan". This solves the so-called lexical gap where users and content producers use different words for the same concept.

What is hybrid search, and why is it important?

Hybrid search combines the precision of keyword search with the understanding of AI-based vector search. This is important because pure AI search can be imprecise – a search for product number "XJ-900" should match exactly, not just "similar" products. With hybrid search, you get the best of both worlds.

Why choose Meilisearch over Elasticsearch?

Meilisearch is written in Rust and optimized for end-user search with response times under 50 ms. It has built-in support for hybrid search (text + vectors) in one solution, and disk-based vector indexing that significantly reduces RAM costs. Elasticsearch is better suited for log analysis and enterprise environments.

What is RAG, and how is it used with search technology?

RAG (Retrieval-Augmented Generation) combines search technology with language models. When a user asks a question, the search engine finds relevant documents via vector search, and these are sent to the AI as context. The result is precise answers based on actual data, without hallucinations.

Does semantic search work with multiple languages?

Yes. Modern embedding models like OpenAI's text-embedding-3 support most languages. Meilisearch also handles multilingual content and can be configured for specific language needs.

How quickly can semantic search be implemented?

With Laravel Scout and Meilisearch, a basic implementation can be in place within 1–2 weeks. More advanced solutions with hybrid search, custom embeddings, and multi-tenancy typically require 4–8 weeks depending on complexity and data volume.

Case Studies

Semantic Search and
Intent-Based Systems

From Keywords to Cognitive Understanding

13 min readSeptember 10, 2025

Semantic Search4 results · 38ms

comfortable running shoes for rainy weather

Nike Air Zoom Pegasus 40Nike

98%

1 899 kr

Adidas Ultraboost LightAdidas

94%

2 299 kr

Hoka Clifton 9 GTXHoka

91%

1 699 kr

Salomon Speedcross 6Salomon

87%

1 499 kr

38msresponse

847Kindexed

94%relevance

About us

Time to build something proper?

Whether you're starting fresh or fixing legacy — we can help.

Semantic Search andIntent-Based Systems

Time to build something proper?

Semantic Search and
Intent-Based Systems