Typesense llms.txt

The actual file

Source Markdown

The saved llms.txt text, displayed as source. The file belongs to its publisher; the live endpoint may change.

Showing the first 18000 characters. Load the complete file to read the remainder.

# Typesense Documentation

> Open-source typo-tolerant search engine. Latest version: 30.2.

## Overview

- [What Is Typesense?](/docs/overview/README.md): Learn what Typesense is, how its open source search engine handles typo-tolerant keyword and vector search, and where to start in the documentation.
- [Typesense Benchmarks](/docs/overview/benchmarks.md): Typesense benchmarks across 2M recipe, 28M book, and 3M product datasets covering RAM use, index time, throughput, and average query latency on small clusters.
- [Comparison with Alternatives](/docs/overview/comparison-with-alternatives.md): How Typesense compares to Elasticsearch, Algolia, and Meilisearch on operability, cost, and developer experience, with a side-by-side feature matrix.
- [Live Demos](/docs/overview/demos.md): Live demo sites showcasing Typesense across recipes, books, geo, vector, semantic, conversational, natural-language, and image search.
- [Typesense Features: Keyword, Vector & AI Search](/docs/overview/features.md): Explore Typesense features for typo-tolerant keyword search, filtering, faceting, geo search, vectors, hybrid search, RAG, JOINs, and clustering.
- [Typesense Use Cases and Examples](/docs/overview/use-cases.md): Explore Typesense use cases including search-as-you-type, faceted browsing, geo and multi-tenant search, semantic search, recommendations, and RAG.
- [What is Typesense?](/docs/overview/what-is-typesense.md): Typesense is an open-source, typo-tolerant search engine optimized for sub-50ms instant search and developer productivity, positioned between Algolia and Elasticsearch.
- [Why Typesense?](/docs/overview/why-typesense.md): Typesense design philosophy: batteries-included defaults, simple operability, and a focus on shrinking time-to-market for a great instant-search experience.

## API Reference (30.2)

- [Typesense API Documentation v30.2](/docs/30.2/api/README.md): Typesense v30.2 API reference covers endpoints, parameters, release changes, and guidance for upgrading or downgrading Typesense Server.
- [Analytics](/docs/30.2/api/analytics-query-suggestions.md): Aggregate search queries and track user interaction events (clicks, conversions, visits) for query suggestions, popularity scoring, and personalization.
- [API Clients](/docs/30.2/api/api-clients.md): Official and community-maintained Typesense client libraries for JavaScript, PHP, Python, Ruby, Dart, Java, Go, Swift, and more, with install links.
- [API errors](/docs/30.2/api/api-errors.md): HTTP response codes returned by the Typesense API and what each 4xx and 5xx status means.
- [API Keys](/docs/30.2/api/api-keys.md): Create, list, and delete Typesense API keys with fine-grained access control scoped per collection, action, record, or field, including scoped search keys.
- [Authentication](/docs/30.2/api/authentication.md): Configure hostnames and API keys to authenticate requests against a Typesense cluster from each official client library.
- [Cluster Operations](/docs/30.2/api/cluster-operations.md): Operate a running Typesense cluster: snapshots, node compaction, vote leadership, health and stats endpoints, slow-request logging, and leader re-election.
- [Typesense Collection Aliases](/docs/30.2/api/collection-alias.md): Use Typesense collection aliases to point a virtual name at a collection, switch indexes without client changes, and support zero-downtime reindexing.
- [Typesense Collections API](/docs/30.2/api/collections.md): Create, list, retrieve, update, and drop Typesense collections. Includes the schema definition format, supported field types, and per-field index options.
- [Conversational Search and RAG API](/docs/30.2/api/conversational-search-rag.md): Build ChatGPT-style conversational Q&A over your indexed data using Typesense built-in RAG pipeline backed by the vector store and pluggable LLM models.
- [Typesense Search Curation API](/docs/30.2/api/curation.md): Curation sets let you pin, hide, or replace documents for specific queries to promote, demote, or substitute results without touching the underlying index.
- [Typesense Documents API](/docs/30.2/api/documents.md): Index, retrieve, update, upsert, delete, import, and export documents in a Typesense collection, including bulk JSONL imports and partial updates.
- [Multi-Search API | Federated Search Across Collections](/docs/30.2/api/federated-multi-search.md): Run multiple search requests in a single HTTP call. Federated mode returns each result independently; union mode merges them into one ranked list.
- [Geo Search API | Filter and Sort by Location](/docs/30.2/api/geosearch.md): Search and filter documents by latitude and longitude using geopoint fields. Supports radius, polygon, and bounding-box queries with distance-based sorting.
- [Image Search API with CLIP Embeddings](/docs/30.2/api/image-search.md): Index and search images using Typesense built-in CLIP model. Search by text description or find visually similar images via embedding-based KNN.
- [Typesense Joins API](/docs/30.2/api/joins.md): Define reference fields between collections and run joined queries with one-to-one, one-to-many, and many-to-many relations across Typesense collections.
- [Natural Language Search API](/docs/30.2/api/natural-language-search.md): Use the Typesense Natural Language Search API to turn plain-English requests into structured query, filter, and sort parameters with an LLM.
- [Typesense Search API Parameters](/docs/30.2/api/search.md): Use the Typesense Search API with query parameters for full-text search, filters, facets, sorting, grouping, typo tolerance, and pagination.
- [Typesense Server Configuration and Ports](/docs/30.2/api/server-configuration.md): Configure Typesense server ports, network binding, storage, peering, TLS, logging, resource limits, and other command-line or config-file options.
- [Stemming](/docs/30.2/api/stemming.md): Enable language-aware stemming so a search for one form of a word also matches its grammatical variations (run/running/ran, company/companies).
- [Stopwords](/docs/30.2/api/stopwords.md): Configure stopword sets, words that Typesense strips from search queries but not from indexed data, scoped per locale and attached to searches.
- [Typesense Synonym Sets API](/docs/30.2/api/synonyms.md): Define one-way and multi-way synonym sets so queries match documents containing equivalent words (e.g. sneaker matches shoe), attached at search time.
- [Typesense Vector Search and Embeddings](/docs/30.2/api/vector-search.md): Index embeddings from any ML model and run nearest-neighbor (KNN) queries for semantic search, recommendations, hybrid search, visual search, and RAG.
- [Voice Query](/docs/30.2/api/voice-search-query.md): Send a base64-encoded audio clip to Typesense for transcription via a configured voice-query model (Whisper), then run the resulting text as a query.

## Guides

- [Typesense Guides and Tutorials](/docs/guide/README.md): Typesense guides provide quick starts, installation steps, search tutorials, API links, reference implementations, and support resources.
- [A/B Testing](/docs/guide/ab-testing.md): Run A/B tests against Typesense search parameters using third-party platforms like Flagsmith, Growthbook, or Optimizely to tune relevance and measure conversion lift.
- [AI Agent Search with Typesense](/docs/guide/ai-agents-typesense.md): Learn why AI agents can use Typesense for keyword, semantic, hybrid, and structured retrieval in one search engine instead of combining separate tools like vector-only stores or plain grep.
- [Building a Search Bar in Angular](/docs/guide/angular-search-bar.md): Build an Angular search UI with Typesense using standalone components, instantsearch.js connectors, and the typesense-instantsearch-adapter, end to end.
- [Build an Astro Search Bar with Typesense](/docs/guide/astro-search-bar.md): Build a typo-tolerant book search interface in an Astro site using Typesense and the typesense-instantsearch-adapter, from project setup to a working UI.
- [Backing Up and Restoring Typesense Data](/docs/guide/backups.md): Safely back up and restore a self-hosted Typesense node by taking a snapshot via the cluster-operations API and archiving the snapshot directory.
- [Boolean Tag Search with Typesense](/docs/guide/boolean-tag-search.md): Implement boolean (AND/OR) search across full-text fields in Typesense via a tag-search pattern, keeping typo tolerance and ranking benefits intact.
- [Build a Search Application with Typesense](/docs/guide/building-a-search-application.md): First-steps walk-through: create a Typesense collection, index a sample books dataset, and run your first searches via a client library.
- [Configure Typesense](/docs/guide/configure-typesense.md): Redirect page to the latest version of the Typesense server configuration reference.
- [Typesense API Keys and Access Control](/docs/guide/data-access-control.md): Manage who can read and write to a Typesense cluster: bootstrap key handling, scoped API keys per collection, action, record or field, and key rotation.
- [Building a Search API with Deno, Hono, PostgreSQL, and Typesense](/docs/guide/deno-hono-search-api.md): Build a Deno search API with Hono, PostgreSQL, and Typesense while keeping the search index synced with source data.
- [Building a Search API with Django and Typesense](/docs/guide/django-search-api.md): Build a Django REST API backed by PostgreSQL as source of truth, kept in sync with Typesense for fast typo-tolerant search served via a proxy endpoint.
- [Deploy a Typesense cluster to Docker Swarm](/docs/guide/docker-swarm-high-availability.md): Deploy a Typesense high-availability cluster across multiple hosts using Docker Swarm services, including init, peer config, and node-join steps.
- [DocSearch for Documentation Sites](/docs/guide/docsearch.md): Add DocSearch to documentation sites with Typesense's scraper and JavaScript search bar, including setup for Docusaurus-style structured content.
- [Full-text Fuzzy Search with DynamoDB and Typesense](/docs/guide/dynamodb-full-text-search.md): Stream DynamoDB changes into Typesense via a Lambda plus DynamoDB Streams pipeline so a DynamoDB table gets full-text, typo-tolerant search via Typesense.
- [Typesense Search FAQs](/docs/guide/faqs.md): Frequently asked questions on Typesense covering keyword search, prefix matching, indexing, ranking, deployment, and operational behavior.
- [Firebase Full-Text Search with Typesense](/docs/guide/firebase-full-text-search.md): Add full-text search to a Firebase / Firestore app by syncing documents into Typesense using the official Firebase Extension or your own trigger.
- [Building a Search API with Go Gin and Typesense](/docs/guide/gin-search-api.md): Build a Go (Gin) REST API backed by PostgreSQL as source of truth, kept in sync with Typesense for fast typo-tolerant search served via a proxy endpoint.
- [Run Typesense in GitHub Actions](/docs/guide/github-actions.md): Run end-to-end tests against a real Typesense instance from inside a GitHub Actions workflow, either via a community action or a manual service step.
- [Typesense High Availability Clusters](/docs/guide/high-availability.md): Run a Raft-backed Typesense cluster for high availability and automatic replication, with quorum sizing guidance and Typesense Cloud managed HA.
- [Install Typesense: Cloud, Docker and Linux](/docs/guide/install-typesense.md): Install Typesense via Typesense Cloud, prebuilt DEB/RPM packages, downloadable binaries for Linux/macOS, or the official Docker image.
- [Installing a client](/docs/guide/installing-a-client.md): Pick and install a Typesense client library: official builds for JavaScript, PHP, Python, Ruby, and community libraries for Go, Java, Rust, Dart, Swift, and more.
- [Building a Search Bar in Kotlin for Android](/docs/guide/kotlin-search-bar.md): Build a native Android book search interface with Kotlin, Typesense, MVVM, coroutines, and a repository-based application architecture.
- [Building an AI-Powered Movie Search Engine with LangChain and Typesense](/docs/guide/langchain-ai-search.md): Build a conversational movie search application with Python, Chainlit, LangChain, Typesense hybrid search, and generated recommendations.
- [Laravel Full-Text Search with Typesense](/docs/guide/laravel-full-text-search.md): Add typo-tolerant full-text search to a Laravel app by wiring Laravel Scout to the Typesense driver, indexing Eloquent models with a few config changes.
- [Multilingual Search with Typesense Locales](/docs/guide/locale.md): Index and search non-English text in Typesense: configure the locale field option, supported languages, tokenization tradeoffs, and best practices.
- [Magento 2 Search with Typesense](/docs/guide/magento2-search.md): Replace Magento 2 default database-backed search with Typesense using the community-built Magento Extension to deliver instant, scaleable product search.
- [Migrate from Algolia to Typesense](/docs/guide/migrating-from-algolia.md): Migration guide for moving from Algolia to Typesense: timeline, document mapping, synonyms and rules export-import, and using the InstantSearch adapter.
- [Full-text Fuzzy Search with MongoDB and Typesense](/docs/guide/mongodb-full-text-search.md): Stream MongoDB documents into Typesense via Change Streams (or the official Node.js CLI) to get typo-tolerant full-text search backed by a MongoDB source.
- [Natural Language Search with Typesense](/docs/guide/natural-language-search.md): Build natural language search with Typesense and an LLM such as Gemini by translating plain-English requests into query, filter, and sort parameters.
- [Building a Search Bar in Next.JS](/docs/guide/next-js-search-bar.md): Build a search interface in a Next.js (React) app using Typesense via instantsearch.js and the typesense-instantsearch-adapter, from setup to deployed UI.
- [Building a Search API with Node.js, Express, Drizzle ORM, and Typesense](/docs/guide/node-drizzle-search-api.md): Build a Node.js search API with Express, PostgreSQL, Drizzle ORM, and Typesense while keeping the search index synced with source data.
- [Building a Search API with Node.js, Express, Prisma, and Typesense](/docs/guide/node-prisma-search-api.md): Build a Node.js search API with Express, PostgreSQL, Prisma, and Typesense while keeping the search index synced with source data.
- [Building a Search API with Node.js, Express, Sequelize, and Typesense](/docs/guide/node-sequelize-search-api.md): Build a Node.js search API with Express, PostgreSQL, Sequelize, and Typesense while keeping the search index synced with source data.
- [Building a Search Bar in Nuxt.js](/docs/guide/nuxt-js-search-bar.md): Build a search interface in a Nuxt.js (Vue) app using Typesense via instantsearch.js and the typesense-instantsearch-adapter, from setup to deployed UI.
- [Model and Organize Typesense Collections](/docs/guide/organizing-collections.md): How to model Typesense data into collections: one-per-entity vs combined, tradeoffs for joins, multi-tenancy, and rebuild or reindex strategies.
- [Personalizing Search Results](/docs/guide/personalization.md): Personalize search rankings per user or user-group in Typesense using popularity ranking, scoped overrides, and ML-based recommendations via vector search.
- [Personalizing Search Results with JOINs](/docs/guide/personalized-search-join.md): Personalize Typesense search results with JOINs by combining products and user-specific data for pricing, filtering, faceting, and sorting.
- [Migrating from PostgreSQL to Typesense for Full-Text Search](/docs/guide/postgres-typesense-migration-fts.md)
- [Query Suggestions](/docs/guide/query-suggestions.md): Build Google/Amazon-style search-as-you-type query suggestions in Typesense using either the analytics aggregation rule or a dedicated suggestions collection.
- [Building a Search Bar in Qwik](/docs/guide/qwik-js-search-bar.md): Build a search interface in a Qwik app using Typesense via instantsearch.js and the typesense-instantsearch-adapter, from setup to deployed UI.
- [Typesense Search Ranking and Relevance](/docs/guide/ranking-and-relevance.md): How Typesense ranks results: text-match scoring (frequency, edit distance, proximity), query_by ordering and weights, field weighting, and tie-breakers.
- [Building a Search Bar in React Native](/docs/guide/react-native-search-bar.md): Build a cross-platform iOS/Android search interface in React Native (Expo) using Typesense for instant, typo-tolerant in-app search.
- [Recommendations](/docs/guide/recommendations.md): Generate item recommendations in Typesense by training a model (e.g. Starspace), storing embeddings, and using vector search for nearest-neighbor lookup.
- [Reference Implementations](/docs/guide/reference-implementations/README.md): Browse Typesense reference implementations with live demos and source code for search, filtering, recommendations, geolocation, and more.
- [Address autocomplete](/docs/guide/reference-implementations/address-autocomplete.md): Build address autocomplete with Typesense using a reference implementation that includes a live interactive demo and reusable source code.
- [AI Image Search](/docs/guide/reference-implementations/ai-image-search.md): Explore image similarity and text-to-image search built with Typesense, Next.js App Router, and the Typesense JavaScript client.
- [Airports Geo Search](/docs/guide/reference-implementations/airports-geo-search.md): Search and filter 78,000 airports on a map with Typesense, Next.js, React InstantSearch, and the Typesense InstantSearch adapter.
- [Books Search](/docs/guide/reference-implementations/books-search.md): Search 28 million OpenLibrary books and filter results by author or subject in this Typesense reference implementation and live demo.
- [Boolean Search](/docs/guide/reference-implementations/boolean-search.md): Implement Boolean search logic in Typesense with a tag-based interface while retaining query-by search across the indexed game data.
- [E-Commerce