How to beginner · 3 min read

Langfuse token usage analytics

Q: Langfuse token usage analytics

Use the langfuse Python SDK to automatically track token usage by wrapping your OpenAI API calls. Initialize Langfuse with your API keys, then decorate or call your LLM functions to collect detailed token usage analytics for monitoring and optimization.

Quick answer

Use the langfuse Python SDK to automatically track token usage by wrapping your OpenAI API calls. Initialize Langfuse with your API keys, then decorate or call your LLM functions to collect detailed token usage analytics for monitoring and optimization.

PREREQUISITES

Python 3.8+
OpenAI API key
Langfuse public and secret keys
pip install langfuse openai>=1.0

Setup

Install the langfuse SDK and openai Python package. Set your environment variables for OpenAI and Langfuse API keys to enable token usage tracking.

bash

pip install langfuse openai>=1.0

Step by step

This example shows how to initialize Langfuse and OpenAI clients, then make a chat completion call with automatic token usage analytics captured by Langfuse.

python

import os
from openai import OpenAI
from langfuse import Langfuse
from langfuse.decorators import observe

# Initialize Langfuse client with public and secret keys
langfuse = Langfuse(
    public_key=os.environ["LANGFUSE_PUBLIC_KEY"],
    secret_key=os.environ["LANGFUSE_SECRET_KEY"],
    host="https://cloud.langfuse.com"
)

# Initialize OpenAI client
client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

# Decorate your function to automatically track token usage
@observe()
def generate_response(prompt: str) -> str:
    response = client.chat.completions.create(
        model="gpt-4o-mini",
        messages=[{"role": "user", "content": prompt}]
    )
    return response.choices[0].message.content

if __name__ == "__main__":
    prompt = "Explain Langfuse token usage analytics."
    answer = generate_response(prompt)
    print("Response:", answer)

output

Response: Langfuse automatically tracks token usage and other metadata for your OpenAI API calls, enabling detailed analytics and cost monitoring.

Common variations

Use @observe() decorator on async functions for asynchronous tracking.
Manually create spans with langfuse.start_span() for custom instrumentation.
Track other LLM providers by wrapping their clients similarly.
Configure Langfuse host URL if using a self-hosted instance.

python

import asyncio
from langfuse.decorators import observe

@observe()
async def async_generate(prompt: str) -> str:
    response = await client.chat.completions.create(
        model="gpt-4o-mini",
        messages=[{"role": "user", "content": prompt}]
    )
    return response.choices[0].message.content

async def main():
    answer = await async_generate("Async Langfuse token usage example.")
    print("Async response:", answer)

if __name__ == "__main__":
    asyncio.run(main())

output

Async response: Langfuse supports async tracking of token usage with the same ease as sync calls.

Troubleshooting

If token usage is not appearing in your Langfuse dashboard, verify your LANGFUSE_PUBLIC_KEY and LANGFUSE_SECRET_KEY environment variables are set correctly.
Ensure your OpenAI client calls are wrapped or decorated with @observe() to enable tracking.
Check network connectivity to https://cloud.langfuse.com or your configured Langfuse host.

✅

Key Takeaways

Use the Langfuse Python SDK to automatically capture token usage analytics for OpenAI API calls.
Decorate your LLM call functions with @observe() to enable seamless tracking.
Set LANGFUSE_PUBLIC_KEY and LANGFUSE_SECRET_KEY environment variables for authentication.
Langfuse supports both synchronous and asynchronous usage tracking.
Troubleshoot by verifying keys, network access, and proper function decoration.

Verified 2026-04 · gpt-4o-mini

Verify ↗