Typesense

What is Typesense?

Typesense is an open source search engine built for fast, typo-tolerant search inside applications. It is made by Typesense and can run self-hosted or as Typesense Cloud, a managed service with dedicated clusters. The problem it solves is relevance and speed: teams get instant results, filters and faceting without building their own indexing and ranking system on top of a general-purpose database.

Key features

• Typo tolerance. Queries still return useful results when users misspell words or type partial terms.
• Instant search. Results update as the user types, which suits search bars, autocomplete and product discovery.
• Faceting and filtering. Results can be narrowed by category, price, tags and other fields, with counts returned alongside matches.
• Vector and hybrid search. Typesense supports vector search and combining keyword relevance with embeddings.
• Geo search. Location fields allow sorting and filtering by distance for local search.
• Curation and synonyms. Teams can define synonyms and override results to promote or hide specific records.
• Managed cloud options. Typesense Cloud offers dedicated clusters, high availability, multiple regions and a search delivery network.

Who is Typesense for?

Developers building search for ecommerce catalogues, documentation sites, marketplaces, SaaS apps and internal tools. It suits teams migrating from a hosted search service that has become expensive, and teams that want to self-host search for data residency or cost reasons. It also fits teams adding semantic or hybrid search to an existing keyword search.

Use cases

• Powering product search and autocomplete in an online store.
• Adding search to documentation, help centres and blogs.
• Building a marketplace search with filters for category, location and price.
• Combining keyword and vector search for a RAG application.
• Replacing an expensive hosted search plan with a right-sized cluster.

Pricing

The open source Typesense server is free to self-host. Typesense Cloud uses dedicated clusters priced by memory, vCPU, region, high availability and optional features such as a search delivery network and GPU acceleration. There are no limits on records or operations, and a free cluster is available to start. Exact prices depend on the selected configuration or are available through the pricing calculator and sales team. See the official Typesense pricing page.

Getting started

1. Create a free cluster at cloud.typesense.org or self-host the open source server.
2. Install a client library for your language.
3. Define a collection schema for the records you want to search.
4. Index your documents and query the collection from your app.

npm install typesense

Alternatives

• Meilisearch is an open source search engine with typo tolerance and instant results.
• Qdrant is a vector database for semantic search and similarity.
• Weaviate is an open source vector database with hybrid search.
• Pinecone is a managed vector database for AI applications.

FAQ

Is Typesense free?

The open source server is free to self-host. Typesense Cloud is paid, with dedicated clusters priced by memory, vCPU and region, and a free cluster available to start.

What is Typesense used for?

Typesense is used to add fast, typo-tolerant search to applications. Common uses include ecommerce product search, documentation search, autocomplete and vector or hybrid search.

Typesense vs Algolia?

Both provide hosted search with typo tolerance and instant results. Typesense can also be self-hosted and prices cloud clusters by memory and vCPU rather than records or operations. The right choice depends on whether you need self-hosting and how your search volume maps to each pricing model.

Does Typesense support vector search?

Yes. Typesense supports vector search and hybrid search that combines vector similarity with keyword relevance.

Can I self-host Typesense?

Yes. The open source server can run on your own infrastructure, including Docker and Kubernetes. Typesense Cloud is the managed option with dedicated clusters and high availability.