Typesense

Typesense is an open-source, developer-friendly search engine designed for high performance and ease of use. It supports both traditional full-text search and vector/semantic search, offering a modern alternative to proprietary solutions like Algolia and Pinecone, as well as open-source engines like Elasticsearch.

Features

• Open-source and self-hostable
• Hybrid search capabilities: Supports both full-text and vector/semantic search
• Vector Search: Nearest-neighbor search using embeddings, with built-in ML models or external APIs (OpenAI, PaLM)
• Search-as-you-type and Autocomplete
• Faceted Navigation and Filtering & Faceting: Aggregate field values, get counts, min, max, avg across records
• Geo Search: Search and sort within a distance from a location or region
• Fuzzy Search: Typo tolerance and automatic correction of spelling mistakes
• Synonyms: Define custom synonym mappings for search queries
• Tunable Ranking: Flexible query-time ranking
• Merchandising: Pin specific records to feature or highlight them
• Multi-tenant API Keys: Store and restrict access to user-specific data in a shared index
• Dynamic Sorting: Sort records on any field dynamically without creating duplicate indices
• Grouping & Distinct: Group results by fields, e.g., aggregate color variations of a product
• Federated Search: Query multiple collections in a single search
• Recommendations and LLM Augmentation
• Easy High Availability: Simple setup for resilient, production-grade deployments
• Intuitive RESTful API
• Platform Integrations: Libraries for many programming languages and integrations with CMS, e-commerce, and dev platforms
• Showcase Demos: Examples include instant search for large datasets (music, books, recipes, code commits), e-commerce, and more

Pricing

• Self-hosted (Open Source): Free
• Typesense Cloud: Paid plans are available for cloud-hosted deployments. (Details not provided in the content above; visit Typesense Cloud for current pricing.)

Category

Vector Database Engines

Tags

open-source, hybrid-search, vector-search, full-text-search