Skill

role-fullstack:search-implementation

Implement search features using Meilisearch, Typesense, Algolia, ElasticSearch, OpenSearch, or PostgreSQL full-text search. Covers indexing strategy with incremental sync via CDC or application hooks, autocomplete/typeahead with debounce, faceted search UI with URL-encoded filter state, relevance tuning (field boosts, typo tolerance, synonyms), and common pitfalls. Use when adding product search, content discovery, autocomplete, or faceted filtering to an application.

Install

npx claudepluginhub rnavarych/alpha-engineer --plugin role-fullstack

Tool Access

This skill is limited to using the following tools:

ReadGrepGlobBash

Preview

- Adding full-text search to a product catalog or content site

Supporting Assets

references/engine-selection-indexing.mdreferences/ui-patterns-relevance.md

SKILL.md

Similar Skills

skill-lookup

Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.

prompts.chat

157.6k

prompt-lookup

Searches prompts.chat for AI prompt templates by keyword or category, retrieves by ID with variable handling, and improves prompts via AI. Use for discovering or enhancing prompts.

prompts.chat

157.6k

android-clean-architecture

Implements Clean Architecture in Android and Kotlin Multiplatform projects: module layouts, dependency rules, UseCases, Repositories, domain models, and data layers with Room, SQLDelight, Ktor.

everything-claude-code

156.2k

Stats

Parent Repo Stars9

Parent Repo Forks0

Last CommitMar 3, 2026

Actions

View Source View Plugin View on GitHub View README

role-fullstack:search-implementation

When to use

Adding full-text search to a product catalog or content site

Building autocomplete or typeahead input

Implementing faceted filtering with sidebar checkboxes and result counts

Syncing a database with a search index (initial bulk load + incremental updates)

Tuning relevance ranking after users complain results are wrong

Choosing between self-hosted (Meilisearch, Typesense) and managed (Algolia)

Core principles

Engine matches scale and budget — pg_trgm for small datasets with no extra infra; Meilisearch/Typesense for fast typo-tolerant search on a budget; Algolia when managed SaaS and rich UI components justify the cost

Denormalize for search — flatten related data into the search document at index time; joins at query time kill latency

Incremental sync, not full re-index — hook into Prisma middleware, DB triggers, or a CDC pipeline; bulk re-indexing on every change is the first thing that breaks under load

Debounce is not optional — 200-300ms debounce on autocomplete input; without it you hit rate limits and burn the user's battery

Encode filters in the URL — active facets belong in useSearchParams; shareable and bookmarkable search state is a product feature, not a nice-to-have

Reference Files

references/engine-selection-indexing.md — engine comparison table with latency and hosting characteristics, four-step indexing strategy, incremental sync options (CDC, hooks, job queue), Meilisearch settings and search code example

references/ui-patterns-relevance.md — debounced useSearch hook with TanStack Query, keepPreviousData pattern, faceted search UI with URL sync, relevance tuning controls (field boosts, typo tolerance, synonyms, stop words, custom ranking), common pitfalls

Search Implementation

When to use

Adding full-text search to a product catalog or content site
Building autocomplete or typeahead input
Implementing faceted filtering with sidebar checkboxes and result counts
Syncing a database with a search index (initial bulk load + incremental updates)
Tuning relevance ranking after users complain results are wrong
Choosing between self-hosted (Meilisearch, Typesense) and managed (Algolia)

Core principles

Engine matches scale and budget — pg_trgm for small datasets with no extra infra; Meilisearch/Typesense for fast typo-tolerant search on a budget; Algolia when managed SaaS and rich UI components justify the cost
Denormalize for search — flatten related data into the search document at index time; joins at query time kill latency
Incremental sync, not full re-index — hook into Prisma middleware, DB triggers, or a CDC pipeline; bulk re-indexing on every change is the first thing that breaks under load
Debounce is not optional — 200-300ms debounce on autocomplete input; without it you hit rate limits and burn the user's battery
Encode filters in the URL — active facets belong in useSearchParams; shareable and bookmarkable search state is a product feature, not a nice-to-have

Reference Files

references/engine-selection-indexing.md — engine comparison table with latency and hosting characteristics, four-step indexing strategy, incremental sync options (CDC, hooks, job queue), Meilisearch settings and search code example
references/ui-patterns-relevance.md — debounced useSearch hook with TanStack Query, keepPreviousData pattern, faceted search UI with URL sync, relevance tuning controls (field boosts, typo tolerance, synonyms, stop words, custom ranking), common pitfalls