Sovereign AI on Hybrid Manager Innovation Release
This documentation covers the current Innovation Release of
EDB Postgres AI. You may also want the docs for the current LTS version.
Hub quick links: Sovereign AI (hub) — AI Factory Hub
What Sovereign AI means in Hybrid Manager
With AI Factory on Hybrid Manager, AI workloads are co-located with your data:
- Your environment. Assistants, model serving, and vector search run directly inside your Hybrid Manager project and Kubernetes cluster.
- Your data. Source documents, embeddings, and conversation history 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. Conversation threads and model logs are observable through the same platform you use for Postgres and analytics.
Connectivity options
- Standard (connected). Most environments run with internet connectivity to pull container images, updates, or libraries. Even when connected, data stays inside your project — unlike SaaS AI, it is not shared with the provider.
- Air-gapped. For high-security deployments, Hybrid Manager can be operated without internet access. You preload model images and registries, and all workloads stay entirely within your controlled cluster.
Contrast with external AI services
- External LLM services (e.g., OpenAI, Anthropic, Cohere): your prompts, documents, and embeddings leave your network and are processed in a third-party environment.
- Hybrid Manager: all retrieval, generation, and storage happen inside your project and infrastructure. Models are served next to your data, ensuring that sensitive content never leaves your control. Air-gapped deployments further guarantee that no external connectivity is required.
Learn more
- Sovereign AI (hub) — principles and patterns.
- AI Factory Hub — deeper guides on assistants, knowledge bases, and model serving.