> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tokenlab.sh/llms.txt
> Use this file to discover all available pages before exploring further.

# Graphiti

> Use TokenLab as an OpenAI-compatible Graphiti LLM provider

## Overview

Graphiti supports OpenAI-compatible `/v1` endpoints through `OpenAIGenericClient`. Point that client at TokenLab and choose a TokenLab model.

<Note>
  **Type**: Knowledge graph memory framework

  **Primary Path**: OpenAI-compatible Chat Completions

  **Support Confidence**: Supported OpenAI-compatible path
</Note>

## Environment

```bash theme={null}
export TOKENLAB_API_KEY="sk-your-tokenlab-key"
```

## LLM Client

```python theme={null}
import os

from graphiti_core.llm_client.config import LLMConfig
from graphiti_core.llm_client.openai_generic_client import OpenAIGenericClient

llm_config = LLMConfig(
    api_key=os.environ["TOKENLAB_API_KEY"],
    model="claude-sonnet-5",
    small_model="gpt-5.4-mini",
    base_url="https://api.tokenlab.sh/v1",
)

llm_client = OpenAIGenericClient(config=llm_config)
```

Pass `llm_client` into your `Graphiti(...)` instance alongside your graph database and embedder configuration.

## Endpoint Notes

Graphiti's OpenAI-compatible path is chat-completions oriented. Use models that reliably return structured JSON for extraction and deduplication tasks.