RAG

Eden RAG grounds models and agents in governed context retrieved from the systems an organization already operates. It combines semantic discovery, lexical and vector retrieval, lineage, provenance, and workload relevance without requiring every source to be copied into a new platform first.

How it works

1. Register the databases, APIs, applications, and other systems that hold

relevant context.

1. Discover schemas, entities, relationships, lineage, and approved samples.
2. Retrieve context using lexical, vector, relationship, provenance, and

operational signals.

1. Apply the requesting actor's permissions and sensitive-data policy before

context reaches a model or agent.

1. Retain route, retrieval, policy, and provenance evidence for the request.

Capabilities

• Schema and entity discovery
• Lexical and vector retrieval
• Relationship and lineage signals
• Permission-aware context assembly
• Sensitive-data redaction
• Agent and AI gateway integration

RAG is currently a preview capability built on Eden's semantic layer, data access controls, and AI gateway.