Graphiti

🚀 Graphiti Project Documentation

Graphiti is a tool for building intelligent applications, supporting multiple large language models (LLMs) and vector databases. By integrating with the Neo4j graph database, it provides powerful knowledge representation and retrieval capabilities.

✨ Features

• Multi-LLM Support: Compatible with mainstream LLM services such as OpenAI, Azure OpenAI, and Google Gemini.
• Vector Embedding: Utilizes text embedding technology for efficient knowledge retrieval.
• Flexible Graph Structure: Allows users to customize nodes and edges to build personalized knowledge graphs.
• Powerful Retrieval Capability: Offers various retrieval options to meet the needs of different application scenarios.

📦 Installation

Install Graphiti

Install using pip:

poetry add graphiti_core  # or
pip install graphiti_core

💻 Usage Examples

Basic Usage

from graphiti_core import Graphiti

# Initialize Graphiti and connect to a local Neo4j instance
graphiti = Graphiti(
    "bolt://localhost:7687",
    "neo4j",
    "password"
)

# Execute a query
response = graphiti.query("Tell me basic information about artificial intelligence.")
print(response)

Advanced Usage

Using Azure OpenAI

from openai import AsyncAzureOpenAI
from graphiti_core import Graphiti
from graphiti_core.llm_client import OpenAIClient
from graphiti_core.embedder.openai import OpenAIEmbedder, OpenAIEmbedderConfig

# Configure Azure OpenAI service
api_key = "<Your API Key>"
api_version = "<Your API Version>"
endpoint = "<Your Azure Endpoint>"

# Initialize the Azure OpenAI client
client = AsyncAzureOpenAI(
    api_key=api_key,
    api_version=api_version,
    endpoint=endpoint
)

# Create a Graphiti instance using Azure OpenAI as the LLM and embedder
graphiti = Graphiti(
    "bolt://localhost:7687",
    "neo4j", 
    "password",
    llm_client=OpenAIClient(client),
    embedder=OpenAIEmbedder(config=OpenAIEmbedderConfig(embedding_model="text-embedding-3-small"), client=client)
)

# Execute a query
response = graphiti.query("Search for articles about machine learning.")
print(response)

Using Google Gemini

from graphiti_core import Graphiti
from graphiti_core.llm_client.gemini_client import GeminiClient, LLMConfig
from graphiti_core.embedder.gemini import GeminiEmbedder, GeminiEmbedderConfig

# Configure the Google API key
api_key = "<Your Google API Key>"

# Initialize a Graphiti instance using Gemini as the LLM and embedder
graphiti = Graphiti(
    "bolt://localhost:7687",
    "neo4j",
    "password",
    llm_client=GeminiClient(config=LLMConfig(api_key=api_key, model="gemini-2.0")),
    embedder= GeminiEmbedder(config=GeminiEmbedderConfig(api_key=api_key))
)

# Execute a query
response = graphiti.query("Explain the basic principles of quantum computing.")
print(response)

📚 Documentation

• Official Guide and API Documentation
• Project Source Code Repository

🔧 Project Status and Roadmap

Current Version

• Stable Version: v1.0.0
• Latest Development Version: v1.1.0-beta

Planned Features

• Support for more LLM providers (e.g., Anthropic)
• Enhanced vector retrieval performance
• Integration of graph data visualization tools

🤝 Contribution Guidelines

The community's contributions are welcome! Please refer to the Developer Documentation to learn how to contribute to the project.

Coding Standards

• Follow the PEP8 coding standard.
• Provide unit test cases.
• Ensure documentation is up-to-date.

📞 Contact Information

For feedback or technical support, please contact us:

• Email: support@graphiti-project.com
• Slack Community: Join the Community

With Graphiti, you can easily build intelligent applications, making complex knowledge management tasks simple and efficient. Start using it now and explore more possibilities!