Other Work Related to Querylight TS

Querylight TS sits in a broader ecosystem of search-related projects. Some of these are directly related, some solve adjacent problems, and some are supporting libraries that make search systems easier to build.

This article covers the current, non-deprecated projects that are most relevant around Querylight TS.

For the Elasticsearch and OpenSearch client line, this intentionally focuses on kt-search and excludes older deprecated repositories such as es-kotlin-client.

Querylight

Repository: jillesvangurp/querylight

This is the Kotlin multiplatform predecessor and sibling project to Querylight TS.

It provides:

• in-memory text indexing
• a small query DSL
• TF-IDF and BM25 ranking
• aggregations
• serialization of index state
• optional vector indexing

If you want to understand where Querylight TS comes from conceptually, this is the most direct related project. The TypeScript port keeps the same general spirit: lightweight, local, structured retrieval without requiring a search server.

kt-search

Repository: jillesvangurp/kt-search

kt-search is the active Kotlin multiplatform client for Elasticsearch and OpenSearch.

It is for:

• talking to Elasticsearch and OpenSearch from Kotlin
• building queries, mappings, templates, and bulk requests with Kotlin DSLs
• running search and indexing logic on JVM, JS, and other Kotlin targets
• operating search clusters from scripts, tools, and applications

This sits at the opposite end of the spectrum from Querylight TS. Querylight TS is for local in-memory search in a browser or Node.js process. kt-search is for integrating with full external search engines.

If Querylight TS answers “how do I ship search with my app without a backend search service?”, kt-search answers “how do I use Elasticsearch or OpenSearch productively from Kotlin?”

Rankquest Core

Repository: jillesvangurp/rankquest-core

rankquest-core is a Kotlin multiplatform relevance evaluation library.

It is for:

• measuring search quality
• scoring ranked result lists with metrics such as precision, recall, MRR, ERR, DCG, and NDCG
• defining portable test cases for search APIs

This is relevant to Querylight TS because retrieval quality is not just about implementing a query engine. At some point you need to evaluate whether ranking is actually good. rankquest-core addresses that measurement problem.

Rankquest Studio

Repository: jillesvangurp/rankquest-studio

rankquest-studio is the browser-based UI built around Rankquest.

It is for:

• creating rated search test cases
• running ranking metrics in a visual UI
• exploring search quality over time
• exporting JSON configurations and judgments

Where Querylight TS helps you build retrieval into an application, Rankquest Studio helps you evaluate whether the results are good enough.

Rankquest CLI

Repository: jillesvangurp/rankquest-cli

rankquest-cli is the command-line companion to Rankquest Studio and Rankquest Core.

It is for:

• running ranking test suites from exported JSON
• integrating relevance testing into local workflows
• enforcing search quality checks in CI

This is relevant if you move from experimenting with search to maintaining it as part of a production workflow.

JsonDsl

Repository: jillesvangurp/json-dsl

json-dsl is a Kotlin multiplatform library for building extensible Kotlin DSLs for JSON and YAML dialects.

It is for:

• creating Kotlin-first DSLs over JSON structures
• keeping DSLs extensible when upstream formats evolve
• supporting plugin-specific fields without redesigning the whole type model

This matters because kt-search uses it for Elasticsearch and OpenSearch DSLs. It is less directly related to Querylight TS itself, but it is part of the broader toolchain around search-oriented developer experience.

GeoGeometry

Repository: jillesvangurp/geogeometry

geogeometry is a Kotlin multiplatform geospatial algorithms library.

It is for:

• geohashes
• GeoJSON manipulation
• polygon and point geometry operations
• coordinate conversions
• shape covering and other geo-indexing helpers

This is relevant to Querylight TS because Querylight TS includes geo search support. The overlap is conceptual rather than implementation-level: both projects care about lightweight search and retrieval primitives, including geospatial ones.

kt-search Logback Appender

Repository: jillesvangurp/kt-search-logback-appender

This project uses kt-search to bulk-index Logback events into Elasticsearch or OpenSearch.

It is for:

• shipping JVM application logs into a search cluster
• structuring log events for search and dashboarding
• making exception and MDC data searchable

This is adjacent rather than directly connected to Querylight TS, but it shows another practical use of the search tooling ecosystem: operational logging and observability.

OpenAI Embeddings Processor

Repository: jillesvangurp/openai-embeddings-processor

This is a small utility project for generating embeddings for vector-search experiments.

It is for:

• preparing embeddings outside the main application
• experimenting with vector retrieval
• exploring semantic search workflows

This connects naturally to Querylight TS because Querylight TS has vector search support and the demo includes semantic “ask the docs” style retrieval.

How these projects fit together

A useful mental model is:

• Querylight TS: local in-process search for TypeScript
• Querylight: the Kotlin sibling and conceptual predecessor
• kt-search: remote search engine integration for Elasticsearch and OpenSearch
• Rankquest: measurement and evaluation of ranking quality
• JsonDsl: DSL infrastructure used by the server-search client ecosystem
• GeoGeometry: geospatial foundations for geo-oriented retrieval work
• embeddings tooling: experiments and support for semantic retrieval

Together these projects cover a wide slice of search work:

• local search
• server-backed search
• relevance evaluation
• geo search
• vector search
• developer tooling and DSL ergonomics