Timescale Logo
backgroundHeroCurvedLine.svg
Coming soon

The fastest ANN search for vector embeddings on PostgreSQL

Timescale Vector improves search speed and recall accuracy for pgvector with a state of the art graph-based index, so you can build scalable LLM applications with PostgreSQL as your vector database.

backgroundPowerChatbotsCurvedLine.svg

Power chatbots, agents and other LLM AI applications with PostgreSQL

Pgvector is an open-source vector database to build AI applications with embeddings data using the PostgreSQL you already know and love. 

Build LLM applications using your data

ChatGPT but for your data

The possibilities are endless: Build customer service bots, chat with your developer documentation, and get easy answers to questions over 100+ page documents.

Retrieval Augmented Generation

Use Retrieval Augmented Generation (RAG) to supplement foundation models with your own data. Augment base large language models like GPT4 from OpenAI, LLaMa, Falcon, Claude from Anthropic, and more.

Long term memory

Create AI agents using PostgreSQL as vector storage for AutoGPT. Leverage popular development frameworks like Langchain and LlamaIndex.

Use the vector database you already know: PostgreSQL

No need to learn a new database or query language

pgvector extends PostgreSQL to handle vector similarity search and storage of embeddings needed for LLM AI applications with reliability and ease.

All the tools you need for vector similarity search

Exact and approximate nearest neighbor search, L2 distance, inner product, and cosine distance, and out-of-the-box support for up to 2000-dimensional vectors.

Yes, it works with that

Leverage the wide ecosystem of PostgreSQL libraries, ORMs, connectors, and tools. From Python and Javascript to the rest of your favorite languages.
backgroundStartInMinutesCurvedLine.svg

Start in minutes,
scale when needed

Timescale is cloud PostgreSQL++

Run pgvector on Timescale’s fully-managed PostgreSQL cloud platform for an easy, cost-effective experience.

Developer conveniences

One-click database forking for testing new models and different embeddings. Read replicas to scale query load. Free consultative support to guide you as you grow.

Secure and private

SOC2 Type II and GDPR compliance. Data encryption at rest and in motion. VPC peering for your Amazon VPC.

Transparent pricing

No “pay per query” or “pay per index”. Decoupled compute and storage for flexible scaling as your needs change.
backgroundResourcesCurvedLine.svg

Resources

Timescale Vector Cookbook

Visit Github repo

Create, store and query OpenAI embeddings with PostgreSQL and pgvector

Try tutorial

What Are ivfflat Indexes in pgvector and How Do They Work

Read explainer

LangChain and pgvector: Up and Running

Try tutorial
Timescale Logo

Get started in minutes

No credit card required • Free for 30 days

Timescale Logo

Subscribe to the Timescale Newsletter

By submitting, you acknowledge Timescale’s Privacy Policy
2023 © Timescale Inc. All rights reserved.