How to beginner · 3 min read

How to get started with Langfuse

Q: How to get started with Langfuse

Use the langfuse Python SDK to integrate AI observability by initializing Langfuse with your API keys from environment variables. Decorate your AI call functions with @observe() to automatically track and log model interactions.

Quick answer

Use the langfuse Python SDK to integrate AI observability by initializing Langfuse with your API keys from environment variables. Decorate your AI call functions with @observe() to automatically track and log model interactions.

PREREQUISITES

Python 3.8+
Langfuse API keys (public and secret)
pip install langfuse

Setup

Install the langfuse Python package and set your API keys as environment variables for secure authentication.

bash

pip install langfuse

Step by step

Initialize the Langfuse client with your public_key and secret_key from environment variables. Use the @observe() decorator on your AI function to automatically trace calls and responses.

python

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

langfuse = Langfuse(
    public_key=os.environ["LANGFUSE_PUBLIC_KEY"],
    secret_key=os.environ["LANGFUSE_SECRET_KEY"],
    host="https://cloud.langfuse.com"
)

@observe()
def call_ai_model(prompt: str) -> str:
    # Simulate AI call (replace with actual AI client call)
    response = f"Response to: {prompt}"
    return response

if __name__ == "__main__":
    result = call_ai_model("What is Langfuse?")
    print(result)

output

Response to: What is Langfuse?

Common variations

You can use Langfuse with different AI models by wrapping their call functions with @observe(). For asynchronous calls, use @observe() on async functions. You can also configure the host parameter if using a self-hosted Langfuse instance.

python

import asyncio
from langfuse.decorators import observe

@observe()
async def async_ai_call(prompt: str) -> str:
    # Simulate async AI call
    await asyncio.sleep(0.1)
    return f"Async response to: {prompt}"

async def main():
    response = await async_ai_call("Async Langfuse example")
    print(response)

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

output

Async response to: Async Langfuse example

Troubleshooting

If you see authentication errors, verify your LANGFUSE_PUBLIC_KEY and LANGFUSE_SECRET_KEY environment variables are set correctly.
If no data appears in the Langfuse dashboard, ensure your decorated functions are actually called and that the Langfuse client is properly initialized.
For network issues, check your internet connection and firewall settings allowing access to https://cloud.langfuse.com.

✅

Key Takeaways

Install the langfuse package and set API keys via environment variables for secure usage.
Use the @observe() decorator to automatically trace AI model calls and responses.
Langfuse supports both synchronous and asynchronous function tracing in Python.
Verify environment variables and network access if observability data does not appear.
Customize the host parameter for self-hosted Langfuse deployments.

Verified 2026-04

Verify ↗