graphiti - Moses Yebei

The Problem

Building knowledge graphs in real time is a pain. Traditional methods often require complete recomputation of the graph for every data change, which is inefficient and slow. This becomes especially burdensome in dynamic environments where AI agents need to quickly adapt to new information without losing historical context.

What This Does

Enter Graphiti. This framework provides a way to continuously integrate user interactions and data into a coherent knowledge graph without the need for a full refresh. You can find the core functionality in the graphiticore directory, where the graphiti.py file defines the main class for graph management.

Graphiti supports a variety of databases as drivers—like Neo4j and FalkorDB—through files such as neo4jdriver.py and falkordbdriver.py. The structure allows for easy querying and updating, which means you can run complex searches with minimal overhead. Want to test your queries? The docker-compose.test.yml file sets up a testing environment that you can spin up with Docker in no time.

Real-World Use

Imagine you're building an AI that manages customer interactions for an e-commerce platform. You could use Graphiti to create a knowledge graph that tracks user preferences and past purchases. Start by integrating data from your customer database into a graph using quickstart/quickstartneo4j.py. Then, when a customer asks, "What were my last three purchases?" you can quickly query the graph for that information without needing to recompute it every time a new purchase is made.

The Bottom Line

Graphiti is solid for anyone needing to create dynamic knowledge graphs without the usual pain points. It's particularly useful for AI applications that require real-time updates. However, if you're working on a small project or something that doesn't need extensive graph capabilities, this might be overkill. Overall, if you’re serious about building interactive AI systems, it’s worth a look.

What the analyser found

Deployment readiness

4/7

Container image
CI pipeline
Lockfile committed
Test suite
README
License
No committed secrets

Composition

200 files

Python152
YAML17
Markdown13
JSON2
TOML1

Docker

Architecture

Top-level areas of the codebase, sized by module count. Arrows show how many imports cross from one area into another.

graph TD g0["graphiti_core 151 modules"] g1["tests 33 modules"] g2["mcp_server 28 modules"] g3["examples 11 modules"] g4["server 11 modules"] g1 -->|97| g0 g3 -->|34| g0 g2 -->|22| g0 g4 -->|8| g0

Module dependencies

The 10 most depended-upon modules of 235, from static import analysis. Red outline marks a module in an import cycle.

graph LR n0_graphiti_core_nodes["nodes"] n1_graphiti_core_driver_driver["driver"] n2_i_core_driver_query_executor["query_executor"] n3_graphiti_core_edges["edges"] n4_graphiti_core_errors["errors"] n5_graphiti_core_helpers["helpers"] n6_models_nodes_node_db_queries["node_db_queries"] n7_models_edges_edge_db_queries["edge_db_queries"] n8_i_core_driver_record_parsers["record_parsers"] n9_phiti_core_llm_client_config["config"] n1_graphiti_core_driver_driver --> n2_i_core_driver_query_executor n8_i_core_driver_record_parsers --> n3_graphiti_core_edges n8_i_core_driver_record_parsers --> n5_graphiti_core_helpers n8_i_core_driver_record_parsers --> n0_graphiti_core_nodes n3_graphiti_core_edges --> n1_graphiti_core_driver_driver n3_graphiti_core_edges --> n4_graphiti_core_errors n3_graphiti_core_edges --> n5_graphiti_core_helpers n3_graphiti_core_edges --> n7_models_edges_edge_db_queries n3_graphiti_core_edges --> n0_graphiti_core_nodes n5_graphiti_core_helpers --> n1_graphiti_core_driver_driver n5_graphiti_core_helpers --> n4_graphiti_core_errors n7_models_edges_edge_db_queries --> n1_graphiti_core_driver_driver n6_models_nodes_node_db_queries --> n1_graphiti_core_driver_driver n0_graphiti_core_nodes --> n1_graphiti_core_driver_driver n0_graphiti_core_nodes --> n4_graphiti_core_errors n0_graphiti_core_nodes --> n5_graphiti_core_helpers n0_graphiti_core_nodes --> n6_models_nodes_node_db_queries classDef cyc stroke:#ef4444,stroke-width:2px; class n0_graphiti_core_nodes,n1_graphiti_core_driver_driver,n3_graphiti_core_edges,n5_graphiti_core_helpers,n6_models_nodes_node_db_queries,n7_models_edges_edge_db_queries cyc;