Skip to main content
Literal AI supports OpenLLMetry and its integrations. You can use the OpenLLMetry integration to log your LLM, AI framework, and RAG pipeline calls to Literal AI. 
from literalai import AsyncLiteralClient
import os
lai = AsyncLiteralClient(api_key=os.environ["LITERAL_API_KEY"])
lai.initialize()
Here is an example of how to log a regular OpenAI chat completion: 
from literalai import AsyncLiteralClient
import os
lai = AsyncLiteralClient(api_key=os.environ["LITERAL_API_KEY"])
lai.initialize()
# Run a regular OpenAI chat completion

@lai.thread()
async def run_openai_chat_completion():
    lai.set_properties(
        tags=["hello", "world"],
        metadata={"foo": 1},
    )

    oai = AsyncOpenAI(api_key=os.environ["OPENAI_API_KEY"])

    await oai.chat.completions.create(
        model="gpt-4o",
        messages=[{ "role": "user", "content": "Hello, how are you?" }]
    )

asyncio.run(run_openai_chat_completion())
Here are examples: Example 1 and Example 2
⌘I