Gen AI Workloads

Gen AI Workloads in EDB Postgres® AI (EDB PG AI) enable you to build trusted AI applications and intelligent assistants — powered by your data, running on your databases and infrastructure, with full governance and auditability.

Rather than relying on opaque external APIs, EDB PG AI helps you build Sovereign AI solutions — where your intellectual property, data, and AI stack are fully in your control:

• Run assistants, search, and AI APIs over your Postgres data and your knowledge bases
• Keep embeddings, documents, and models in your data estate (Postgres or object storage)
• Build, deploy, and govern Gen AI apps on Kubernetes and Hybrid Manager (EDB), or Postgres workloads with the appropriate licensing
• Audit and trace AI interactions and data flows

Why Sovereign AI?

Sovereign AI means:

• No lock-in to external APIs
• No sensitive data leaving your data estate
• Full transparency over embeddings, search behavior, and model usage
• Control over where inference happens (on your clusters or governed cloud)

EDB PG AI makes this possible — combining:

• Postgres-native vector search
• Flexible Knowledge Bases with retrievers
• Agent Studio and Gen AI Builder to create and manage AI workloads
• Modular pipelines and AI Accelerator integrations

This allows you to run Gen AI Workloads with trust — across Hybrid Manager, on-prem, or licensed EDB workloads.

Where Gen AI Workloads Run

Currently, Gen AI Workloads run within the Hybrid Manager™ platform from EDB. This provides the underlying architecture for:

• Gen AI Builder
• Pipelines
• Vector Engine
• Agent Studio
• AI model serving and pipelines orchestration

For Hybrid Manager-specific architecture and configuration guidance, see: Gen AI Workloads in Hybrid Manager

EDB Postgres AI offerings are always focused on Sovereign AI, with your data estate at the center.

What You Can Build

With Gen AI Workloads, you can:

• Create Enterprise Knowledge Assistants over your Postgres and unstructured data
• Power RAG-based search portals across structured + unstructured content
• Provide Gen AI APIs over your governed Knowledge Bases
• Run secure search and similarity retrieval over sensitive data
• Automate workflows with intelligent agents that use your systems
• Drive composable AI pipelines from Postgres and object storage

See our Use Cases and Solutions for detailed patterns.

Core Components

Explore the components that enable Gen AI Workloads:

Agent Studio

Design, deploy, and manage intelligent Gen AI Assistants:

• Define assistant behavior with rulesets, tools, and retrievers
• Use Knowledge Bases and Pipelines as source data
• Expose assistants via API, chat UI, or internal tools
• Govern behavior and interactions with full auditability

Gen AI Builder

Prepare, govern, and orchestrate your AI knowledge stack:

• Build and manage Knowledge Bases
• Define and manage retrievers to unify structured and unstructured data
• Connect and manage Data Lakes with Pipelines
• Govern data preparation, embedding, and search flow

Vector Engine

Extend Postgres with native vector types and similarity search:

• Store embeddings in Postgres or object storage
• Enable high-performance vector search with full SQL integration
• Drive RAG patterns with full transparency

Where to Start

To begin building Gen AI Workloads:

• Follow the AI Factory 101 Learning Path
• Explore the Explained Concepts — Knowledge Bases, retrievers, Pipelines, and more
• Build your first Assistant in Agent Studio
• Deploy Knowledge Bases using Gen AI Builder

Build Trusted AI with EDB PG AI

EDB Postgres AI lets you deliver trusted Gen AI Workloads — built on your data, in your databases, with transparency.

• Use your models or models under your governance
• Run in your environment — Hybrid Manager (current), with future support for other EDB PG AI workloads
• Keep your Knowledge Bases and embeddings in your data estate
• Provide full auditability for compliance and trust

Build intelligent applications — with your data, on your Postgres estate — today.