Agent Factory in Hybrid Manager v1.4.0 (LTS)

The Agent Factory workload in Hybrid Manager brings scalable AI and machine learning capabilities to your Hybrid Manager platform. It enables you to operationalize AI across your Hybrid Manager–managed clusters and data — with deep integration across Postgres, vector search, model serving, and AI pipelines.

With Agent Factory in Hybrid Manager, you can:

• Build AI flows and agents with Langflow for internal or external-facing use.
• Serve AI models at scale with integrated KServe-powered inferencing and GPU acceleration.
• Create knowledge bases and perform Retrieval-Augmented Generation (RAG) with vector search.
• Manage your model library and deploy trusted models under Hybrid Manager governance.
• Integrate AI features directly into your applications and data pipelines.

Example AI solutions you can build

Agent Factory in Hybrid Manager enables solutions across many domains:

• Enterprise search and knowledge assistants Build RAG-based assistants that integrate with corporate documents, databases, and intranet content.

• Customer support chatbots Deploy assistants powered by your own data and domain-specific models, combining conversational AI with Knowledge Bases.

• AI-driven data apps Expose AI-powered endpoints for apps — semantic search, recommendations, similarity search, or natural language querying.

• Operational AI for internal tools Build AI agents and tools to assist with DevOps, customer success, HR automation, sales enablement, and more.

• Domain-specific model serving Serve proprietary or fine-tuned models (LLMs, embedding models, ranking models) as scalable inference services within your Hybrid Manager cluster.

You can start small — with a single assistant or model endpoint — and scale to full AI-powered applications using the Agent Factory architecture.

Agent Factory in Hybrid Manager workloads

Hybrid Manager supports a full set of Agent Factory capabilities, tightly integrated into its control plane:

Build AI flows

• Langflow — Build and deploy AI pipelines and flows using EDB components connected to HM-managed models, knowledge bases, and Postgres clusters
• Pipeline Designer — Create AIDB data pipelines and knowledge bases through a no-code visual interface

Model management and serving

• Model Library — Manage and govern your model assets
• Model Serving — Deploy and scale model inference services on Kubernetes
• GPU resource management — Configure and allocate GPU capacity for serving

How Agent Factory fits within Hybrid Manager

Hybrid Manager is the control plane for your databases, analytics, and AI workloads. Agent Factory complements other HM capabilities:

• Databases: Provision and manage Postgres clusters; Agent Factory connects to these for vector search, RAG, and application endpoints.
• Analytics: Offload data to Iceberg/Delta and query via SQL; Agent Factory’s pipelines and Vector Engine integrate with these datasets for retrieval.
• Operations: Use HM observability and governance to monitor AI flows, pipelines, and model serving alongside databases and analytics.

Prerequisites for Agent Factory in HM:

• A Hybrid Manager project with Agent Factory enabled and required entitlements for Langflow, Pipeline Designer, and Model Serving.
• Access to GPU resources where needed for serving. See GPU resource management.

Sovereign AI in Hybrid Manager

Agent Factory on Hybrid Manager runs AI workloads co-located with your data:

• Your environment. Model serving, pipelines, and vector search run directly inside your Hybrid Manager project and Kubernetes cluster.
• Your data. Source documents, embeddings, and inference results never leave your environment unless you explicitly configure it.
• Your models. Model images are curated in the Model Library and served from endpoints inside your cluster.
• Your visibility. Model logs and AI workload metrics are observable through the same platform you use for Postgres and analytics.

This is distinct from external AI services, where prompts, documents, and embeddings leave your network and are processed in a third-party environment. With Hybrid Manager, all retrieval, generation, and storage happen inside your project and infrastructure.

For high-security deployments, Hybrid Manager can operate fully air-gapped — model images are preloaded, and all workloads stay entirely within your controlled cluster. See Prerequisites for air-gapped setup.

Get started

• New to Agent Factory on HM? Start here: Getting started
• Review environment requirements: Prerequisites
• Explore real-world patterns: Use cases and personas
• Understand the architecture: Architecture

Getting started

Understand

Model management

Build with AI

Operate & Observe