Core components
note
Knowledge Hub is an Early Access release. Refer to the Early Access disclaimer for details.
Knowledge Hub organizes your data using three core components: repositories, Knowledge Bases, and Sources.
What each component does
| Component | What it is | Learn more |
|---|---|---|
| Repository | The top-level storage container. It holds one or more Knowledge Bases and controls the infrastructure on which vector data is stored. | Repositories |
| Knowledge Base | The core indexing unit. It stores the vector embeddings generated from one or more Sources and serves as the retrieval target for AI agents. | Knowledge bases |
| Source | A single data connection between an external system and a Knowledge Base, created automatically when a source-to-target flow deploys in Data Integration. | Sources |
Next steps
- Refer to Repositories to learn what a repository controls and when to use more than one.
- Refer to Knowledge bases to learn how a Knowledge Base indexes and serves content.
- Refer to Sources to learn how data connections are created and managed.
- Refer to Asset states reference for the full state model across all three components.