Graphiti

Overview

Graphiti is a framework for building and querying temporal context graphs for AI agents. Unlike static knowledge graphs, it tracks how facts change over time, maintains provenance to source data, and supports both prescribed and learned ontology. It continuously integrates user interactions, structured and unstructured enterprise data, and external information into a coherent, queryable graph.

Key Features

• Temporal Fact Management: Facts have validity windows; old facts are invalidated not deleted. Episodes & Provenance: Every entity and relationship traces back to raw episodes. Prescribed & Learned Ontology: Define entity and edge types via Pydantic models or let structure

Use Cases

• Build temporal knowledge graphs that evolve with every interaction. Represent rich, structured context instead of flat document chunks. Query across time, meaning, and relationships with hybrid retrieval.

Getting Started

• Install with pip install graphiti-core. Requires Python 3.10+, a graph database (Neo4j, FalkorDB, Amazon Neptune, or Kuzu), and an OpenAI API key. See the Quickstart example for connecting, initializing indices, adding episodes, and searching.

Deployment & Requirements

• Python 3.10 or higher; Neo4j 5.26 / FalkorDB 1.1.2 / Amazon Neptune Database Cluster or Neptune Analytics Graph + Amazon OpenSearch Serverless collection / Kuzu 0.11.2 (deprecated); OpenAI API key (defaults to OpenAI for LLM inference and embedding). Optional: Google Gemini,

Before You Adopt

• License: Apache-2.0. Review its terms before using, modifying, or distributing the project.
• Graphiti works best with LLM services that support structured output; other services may cause incorrect schemas and ingestion failures.