Context engineering platform for AI agents with temporal knowledge graphs and sub-200ms retrieval
- Stars4.7k
- Forks634
- Open Issues23
Apache-2.0
- Python
- Go
- Jupyter Notebook

About Zep
Zep is an end-to-end context engineering platform for AI agents. It addresses the agent context problem by assembling relationship-aware context from chat history, business data, documents, and app events, then delivering the right information at the right time with sub-200ms latency.
Zep ingests messages, business data, and events as they occur. Its Graph RAG layer extracts relationships and maintains a temporal knowledge graph, so context can reflect facts, relationships, and how they have changed over time. It returns pre-formatted context blocks optimized for an LLM.
Zep is offered through Zep Cloud. The legacy Community Edition is deprecated and no longer supported, with its code moved to a legacy folder. Zep is powered by Graphiti, an open-source temporal knowledge graph framework for relationship-aware context retrieval.
Key features
- Ingests chat messages, business data, documents, and app events
- Graph RAG extracts relationships into a temporal knowledge graph
- Retrieves pre-formatted context blocks optimized for LLMs
- Tracks fact validity with valid_at and invalid_at dates
- Relationship-aware retrieval for facts and changing context
Details
- First released
- 2023
- Latency
- Sub-200ms retrieval
- Deployment
- Zep Cloud
- Graph
- Temporal knowledge graph
- Inputs
- Chat · Data · Docs · Events
- Open source
- Graphiti framework
