Skip to main content

Vector Search

Tigris makes it easy to build high-performance vector search applications. It is a fully managed cloud-native database that allows you store and index documents and vector embeddings for fast and scalable vector search.

Tigris provides the following features:

  • High performance: Tigris is built from the ground up to provide fast vector search. It is optimized for low-latency vector search and high throughput.
  • Real-time: Tigris provides real-time updates to the index. You can add, update, and delete documents and embeddings in real-time.
  • Hybrid search: Tigris supports hybrid search, which allows you to combine vector search with attribute filtering for more relevant search results.
  • Fully managed: Tigris is a fully managed cloud-native database. You can get started with Tigris in minutes and scale as needed with ease.

Vector Search Workflow

A typical vector search workflow with Tigris involves the following steps:

  1. Generate embeddings for your documents.
  2. Index your documents and embeddings in Tigris.
  3. Generate embeddings for your query.
  4. Search for similar documents using the query embeddings.

Vector Search Workflow

Use Cases

Tigris is a general-purpose vector search engine. It can be used for a variety of use cases, including:

  • Similarity search: You can generate vector embeddings from your documents and index and search through the vectors using Tigris to power your similarity search application.
  • Product recommendations: You can store product information and embeddings in Tigris and use it to power your product recommendations for e-commerce.
  • Hybrind search: You can combine vector search with attribute filtering in Tigris for more relevant search results.
  • Image search: You can generate vector embeddings for image data and index them in Tigris to power your image search application.
  • Generative question answering: You can use Tigris to index relevant context in the form of vector embeddings and use it to suppliment a generative model to generate more accurate answers.

Getting Started

Continue to one of the following guides to learn how to get started with Vector Search using Tigris.