It searches the meaning, not just the word. Even with typos…
MediaPress V6 doesn't limit on-site search to classic keyword matching. It uses a hybrid technology that blends vector-based semantic search with keyword search: even when a visitor doesn't know the exact word, it understands what they mean and surfaces the most relevant content.
How you configure hybrid search in the panel
Vector + keyword, together
Hybrid scoring
Meaning-based vector similarity and classic keyword matching are blended into a single result. The best of both worlds: conceptual closeness and exact-term precision. Results are scored from most to least relevant.
Semantic search
Your content is turned into numeric meaning representations (embeddings). When a visitor types "contact form," your "Get in touch" page matches even without those exact words — because the system understands intent, not just wording.
Cross-lingual search
The semantic layer works across languages: a query written in one language can match relevant content by meaning. On multilingual sites, visitors reach the right content even when they search in their own language.
Publish and it's indexed automatically
When a page, blog post or product is published, it's automatically (re)indexed for search. No manual indexing chores; only published content is searchable and drafts never leak in.
Fine-tune from the panel
The search provider, model, title/meta/body/tag score weights, similarity threshold and result limit are all set from the panel. Shape search behavior to your content and audience without writing code.
Protected against abuse
The public search endpoint is protected with per-IP rate limiting and repeated queries are cached. Search stays fast while resources stay safe from abuse.