Solutions built with Zep | Zep Documentation

Zep is the unified context layer for enterprise data. It combines business data, documents, and conversations into shared, governed context.

The three solution categories use the same Context Lake and Context Graph lifecycle. The correct solution depends on the subject of the context and the agents that can retrieve it.

Enterprise context graphs

Unify knowledge about projects, products, operations, and business domains for agents across your organization.

Agent memory

Give an agent context from past conversations, user activity, and changing preferences.

Customer and account context

Combine account records, conversations, and events so an agent can understand each customer’s situation.

Choose how the agent retrieves context

Each solution supports one-call grounding and an agent with Zep tools. For one-call grounding, a user graph uses thread.get_user_context, and a shared Context Graph uses graph.search with scope="auto". For complex questions and tasks, the agent uses Zep through tools. See Build an Agent with Zep.

Use more than one solution

One task can require context from multiple graph scopes. For example, a customer support agent can use account events, a user’s previous conversations, and an enterprise support policy.

Keep each data set in the graph that matches its subject and access boundary. Retrieve only the context that the caller can access. Then combine the selected results in the application.

Use Get started with Zep for a graph-scope comparison. Use Architecture patterns for ingestion and retrieval decisions.