How do you implement search functionality in your apps?

Search ranges from simple to sophisticated depending on your needs.

Basic search uses SQL LIKE or ILIKE: WHERE name ILIKE '%query%'. It works for small datasets and simple matching but doesn't handle typos, relevance ranking, or synonyms. Good enough for admin panels or simple filters.

Database full-text search (PostgreSQL tsvector, MySQL FULLTEXT) is the next step. It tokenizes text, handles stemming (finding "run" matches "running"), and ranks results by relevance. No external services needed.

Dedicated search engines (Elasticsearch, Meilisearch, Typesense, Algolia) offer superior features: typo tolerance, faceted filtering, instant results, and relevance tuning. They index your data separately and require syncing when data changes.

Algolia and Typesense are simpler to set up than Elasticsearch. Algolia is hosted (no infrastructure), Typesense is lightweight enough to self-host easily.

Implementation pattern: sync relevant data to search engine on create/update/delete (webhooks or queue jobs). Frontend sends queries to search API, which returns IDs. Optionally fetch full records from your database using those IDs.

For auto-complete, debounce client queries (don't search on every keystroke), return results fast (under 100ms feels instant), and highlight matching text in results.

Start simple. Database search handles most use cases. Add a dedicated engine when you need typo tolerance, complex filtering, or performance at scale.