Graphiti
Graphiti is Zep's open-source temporal knowledge graph framework that ingests agent conversations + structured data, extracts facts with timestamps, and supports time-aware queries over the graph.
Graphiti (open-sourced by Zep in 2024) is a knowledge graph framework purpose-built for agent memory. It ingests conversations + structured data, extracts entities + relationships + facts via LLMs, and timestamps each fact for temporal querying ("what did this user prefer last week vs this week?"). Used as the backbone of Zep's production memory layer and adopted standalone by teams building custom agent memory. Notable feature: explicit handling of fact invalidation when new information contradicts old.
When to use graphiti
- Production agent memory with temporal facts.
- Knowledge graphs for AI agents.
Common mistakes
- Treating Graphiti as a generic KG framework — it's purpose-built for agent memory shapes.
FAQ
What is graphiti?
Graphiti is Zep's open-source temporal knowledge graph framework that ingests agent conversations + structured data, extracts facts with timestamps, and supports time-aware queries over the graph.
When should I use graphiti?
Production agent memory with temporal facts. Knowledge graphs for AI agents.
What are the most common mistakes with graphiti?
Treating Graphiti as a generic KG framework — it's purpose-built for agent memory shapes.
Related terms
- Graph RAG — Graph RAG builds a knowledge graph from the corpus during ingestion — entities, relationships, facts — and retrieves via graph traversal alongside vector search, improving recall on relational queries.
- Stateful agent — A stateful agent persists state — memory, learned facts, long-running context — across sessions, in contrast to stateless agents that start fresh on every conversation.
Last updated: 2026-06-01. Raw markdown: https://promtable.com/glossary/graphiti.md.