Skip to main content

Use Cases

hlquery is built for teams that need fast retrieval, relevance controls, vector search, and straightforward operations without taking on a large search stack.

E-Commerce and Retail​

Improve product discovery and conversion.

  • Typo tolerance: Keep useful results available when users misspell product names.
  • Filtering and faceting: Let users narrow by category, price, availability, brand, or other schema fields.
  • Synonyms and stopwords: Normalize common product language across global and collection-specific dictionaries.
  • Overrides and aliases: Promote seasonal products and move traffic between indexed collection versions.
  • Vector and hybrid search: Support semantic product discovery when exact terms are not enough.

Content and Publishing​

Prioritize fresh, relevant content across articles, guides, and media libraries.

  • Multi-field search: Query titles, summaries, tags, and body text.
  • Facets: Group results by author, category, date, or content type.
  • Aliases: Publish a new collection version behind a stable name.
  • Global search: Query across collection targets when the application does not know the exact collection up front.

Documentation and Knowledge Bases​

Make large documentation sets easier to search.

  • Lexical search: Return exact matches for APIs, options, commands, and error strings.
  • Vector search: Retrieve conceptually similar answers for natural-language questions.
  • Hybrid search: Blend exact technical terms with semantic recall.
  • Version isolation: Store each product or documentation version in its own collection.

Operational Logs and Analytics​

Search operational data during incidents and investigations.

  • Fast writes: Use bulk import and durable RocksDB-backed storage.
  • Filters: Narrow by service, severity, host, tenant, or time-like fields.
  • Export: Extract matching documents for offline analysis.
  • Metrics endpoints: Monitor health, readiness, counters, storage, and runtime stats.

AI and Customer Support​

Use hlquery as a retrieval layer for support tools and assistant workflows.

  • Vector search: Retrieve semantically similar tickets, articles, or product entries.
  • Hybrid search: Keep exact identifiers and semantic language in the same workflow.
  • Access control: Use scoped API keys and embedded filters for tenant-aware retrieval.

Run search close to users or inside smaller deployments.

  • C++ runtime: Lower baseline overhead than many JVM-based stacks.
  • HTTP/JSON API: Simple integration from local services and scripts.
  • Configurable modules: Enable optional modules only when the deployment needs them.
  • CLI and talk REPL: Operate and debug local servers without a separate control plane.

When hlquery Fits​

hlquery is a good fit when you need:

  • REST-first indexing and search.
  • Collection-based schemas.
  • Lexical, vector, and hybrid retrieval in one service.
  • Operational endpoints for health, readiness, metrics, storage, and maintenance.
  • Scoped keys, users, and optional IP filtering.
  • Local development that can move to service deployment without changing API shape.

When to Validate Carefully​

Validate with your own workload when:

  • Your vector dataset is large enough that exact scan cost matters.
  • You need strict P99 latency guarantees.
  • You plan to run heavy write and search traffic at the same time.
  • You depend on optional modules or LLM-backed behavior.
  • You need a specific ranking behavior across many domain-specific query types.