> For the complete documentation index, see [llms.txt](https://docs.boldminded.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.boldminded.com/dexter/docs/dexter-vs-pro-search.md).

# Dexter vs Pro Search

| Feature                   | Dexter                                                                     | Pro Search                                                     |
| ------------------------- | -------------------------------------------------------------------------- | -------------------------------------------------------------- |
| **Search Engine**         | Uses Meilisearch or Algolia (built for search)                             | Uses MySQL full-text (built for storage)                       |
| **Search Approach**       | Context-based, relevance-driven, semantic-style ranking                    | Keyword-only text matching                                     |
| **Speed & Performance**   | Lightning-fast, optimized for large datasets & instant results             | Good performance on small/medium data, slower as content grows |
| **Scalability**           | Designed for heavy traffic & large content sets                            | Limited by MySQL indexing & server load                        |
| **Has JSON API**          | Yes                                                                        | No                                                             |
| **Content Types Indexed** | Entries, categories, members, files, documents, images, custom sources     | Primarily channel entries & categories                         |
| **Filtering & Faceting**  | Advanced faceted filters, live updating, instant results                   | Basic filters (category, ranges, relationships)                |
| **Relevance & Ranking**   | Modern ranking rules, typo tolerance, fuzzy matching, contextual weighting | Field weighting only, limited full-text ranking                |
| **Search Experience**     | Ideal for autocomplete, instant search, headless/JS UIs                    | Traditional keyword search pages                               |
| **Customization**         | Flexible indexing pipelines, transform/enrich data, API-driven             | Collection settings, field weighting, basic API                |
| **Infrastructure**        | Requires Meilisearch or Algolia (self-hosted or SaaS)                      | No external services; runs entirely in EE                      |
| **Indexing**              | Real-time + queued indexing; designed for continuous updates               | Manual or triggered index rebuilds                             |
| **Multimedia Support**    | Extracts text from PDFs, documents, metadata, AI image descriptions        | No document or image indexing                                  |
| **Best For**              | Large sites, modern UX, multi-content search, fast performance             | Smaller sites, simple keyword search needs                     |