AI FACTORY
GenAI Orchestration
Turn all your data into AI-ready assets while simplifying management with AI Pipeline, Vector Engine, and Model Serving.
Key benefits
Accelerate GenAI deployment
Develop faster without manually building and rebuilding AI pipelines. Just use five lines of code in your trusted Postgres® environment to enable vector data capabilities, automated data preparation and embedding, and seamless language model integration.
Ensure accurate GenAI inferencing
Deploy trustworthy GenAI applications with always-current knowledge bases and reranking capabilities for more accurate semantic search. EDB Postgres AI (EDB PG AI) Factory instantly syncs embeddings with source data changes, eliminating the risk of stale or missing records.
Power high-performance, sovereign GenAI applications
Create enterprise-grade GenAI with vector similarity and semantic search that's 4.22x faster than specialized vector databases. Seamlessly bridge proprietary data with swappable, multimodal models delivered to your enterprise infrastructure—or bring your own custom models.
Key features
AI Pipeline
Preparers
Turn all your data into AI-ready assets with preparers that perform common preprocessing steps on source data—including cleaning, chunking, embedding, and adding metadata—so they can be embedded and indexed for use by knowledge bases.
Auto-embedding
Build responsive, accurate GenAI applications with automatic generation and refresh of embeddings from your source data. This automatic processing handles all data changes on demand in optimized batches, eliminating the need for manual triggers or concerns about stale data.
Intelligent knowledge bases
Simplify data management with turnkey knowledge bases: repositories of AI-ready data pulled from easily swappable storage locations like Postgres and Amazon S3-compatible object storage. These include all similarity calculations to make the data semantically searchable—enriched with reranking capabilities for greater accuracy.
Vector Engine
Vector data store
Center your AI data strategy on familiar, trusted Postgres with native vector storage—ensuring complete AI sovereignty and eliminating data movement to external vendors.
Semantic search
Power high-performance GenAI applications with semantic search that’s 4.22x faster than specialized vector databases—powered by open source pgvector and intelligent knowledge bases that handle vector similarity calculations for you.
Model Serving
Swappable models
Seamlessly swap between models to meet evolving business needs without infrastructure changes and eliminate vendor lock-in with flexible model deployment—in the cloud, on-prem, or on hardware like NVIDIA and Supermicro with intelligent scaling that optimizes resources automatically.
Multimodal model support
Quickly build cognitive AI applications that can process and understand various data types with human-like understanding—such as text, images, audio, and more—with support for multimodal models from leading providers like Hugging Face and NVIDIA NIM.
Build or bring your own models
Tailor your AI capabilities precisely to your business requirements. Bring your own custom models to EDB PG AI or leverage Red Hat OpenShift AI integrations to train new purpose-built models.
AI Factory Architecture
Documentation and Downloads
Ready to dig in? Access technical product documentation, release notes, product information, and software downloads from EDB repositories.
Resources
EDB Postgres AI Factory: GenAI Builder
EDB Postgres AI Factory: Agent Studio
Sovereign AI Use Case
Cognitive AI Use Case
Virtual Expert Use Case
ISV Webinar
AI Factory
EDB PG AI provides a complete GenAI inferencing platform in AI Factory—with code when you need it and UI when you don’t—enabling rapid deployment of production-ready, sovereign GenAI applications and agents in days instead of months.
EDB Postgres AI
EDB PG AI is an intelligent platform for transactional, analytical, and AI workloads that meets you wherever you are — on premises or on any cloud, anywhere, and on appliances of choice. Get access to the most flexible, open, and extensible general-purpose database, break data silos, and launch new AI initiatives with the highest assurance of security, compliance, and availability.