EDB Postgres AI - AI Accelerator - Overview

What is the AI Accelerator?

AI Accelerator simplifies AI data management by combining vector search from open source PGvector with Pipelines, an extension for AI data pipeline automation. It handles vector embedding operations for structured and unstructured data, simplifying similarity searches and natural language querying. It streamlines development of AI-powered applications by providing retrieval of AI-ready data in a familiar Postgres environment.

Features

Vector toolkit

The EDB Pipelines extension enables automated vector embedding, storage, and retrieval workflows, and it comes preloaded with PGvector for seamless management of vector data in Postgres. This enables developers to build complex GenAI functionality using SQL commands in the familiar Postgres environment with just 5 lines of code instead of 130+.

Swappable configurations

Easily switch between models, integrate multiple data modalities, and select from a variety of storage locations in the cloud or on prem. This flexibility enables tailored performance and efficient data management that adapts to any need across your business.

Automated Pipelines

Simplify data processing with Automated Pipelines for fetching data from Postgres or object storage. Pipelines can perform preprocessing steps directly with the AI Accelerator to make data suitable for embedding generation. This capability eliminates the need for many external ETL workflows and dependencies. Pipelines can also generate vector embeddings as new data is ingested, triggering updates to embeddings when source data changes. This means data is always up to date for query and retrieval without tedious maintenance.

GenAI query engine

Leverage semantic search across text and images that’s 4.22X faster¹ than purpose-built vector databases and an intelligent retriever that abstracts away the complexities of vector similarity calculations. These capabilities transform your Postgres database into a powerful GenAI search engine across multiple storage locations.

Use cases

• Sovereign AI: Your controlled, adaptable AI platform. Secure, flexible, and cost-effective where your data lives.
• Virtual expert: Natural-language AI agents (for example, chatbots, copilots) to improve workforce and customer outcomes.
• Cognitive AI: Intelligent search, analysis, and recommendation across diverse data types.