Tools tagged with "Search"

Tools tagged "search" provide full-text or faceted search over a dataset: documents, product catalogs, log lines, code, or application records. Open source search engines index content so users can query it with typo tolerance, ranking, filters, and highlighting, instead of relying on slow SQL LIKE queries. Some are broad-purpose engines (Elasticsearch, OpenSearch, Meilisearch, Typesense) meant to sit behind any application; others are narrower, built for a specific use case like site search widgets, log search, or vector/semantic search over embeddings for AI applications.

Common use cases include e-commerce product search with filters and autocomplete, documentation and knowledge base search, internal admin tools searching across records, log and event search for observability, and retrieval for AI chat applications backed by vector similarity. Search infrastructure is also embedded inside larger platforms: a CMS, a helpdesk, or a CRM often ships its own search index rather than exposing one directly.

When choosing a search tool, check indexing speed and how it handles updates (near-real-time vs. batch reindexing), relevance tuning options (typo tolerance, synonyms, custom ranking rules), memory and disk footprint at your expected data size, whether it supports faceted filtering and geo queries, and whether it needs a separate vector index for embeddings or handles hybrid keyword-plus-vector search natively. Operational overhead matters too: Elasticsearch and OpenSearch are powerful but need cluster management, while Meilisearch and Typesense aim for simpler single-binary deployment.

Self-hosted search tools replace paid services like Algolia, Elastic Cloud, and Amazon Kendra, along with in-house search built on raw SQL that struggles past a few hundred thousand rows. Running search yourself avoids per-query or per-record pricing and keeps query logs and indexed content inside your own infrastructure, at the cost of managing uptime and reindexing pipelines yourself.