Langfuse Instrumentor

Installation

$ pip install respan-instrumentation-langfuse

Classes

LangfuseInstrumentor

OpenTelemetry-compliant instrumentor that patches Langfuse’s OTLP exporter to redirect traces to Respan.

1 from respan_instrumentation_langfuse import LangfuseInstrumentor

Methods

Method	Description
`instrument(**kwargs)`	Enable instrumentation. Must be called BEFORE importing Langfuse.
`uninstrument(**kwargs)`	Disable instrumentation.

Important

LangfuseInstrumentor().instrument() must be called before importing langfuse. The instrumentor patches the OTLP exporter at import time.

Usage

1 import os
2 from respan_instrumentation_langfuse import LangfuseInstrumentor
3 
4 # Set credentials
5 os.environ["RESPAN_API_KEY"] = "your-api-key"
6 os.environ["LANGFUSE_PUBLIC_KEY"] = "your-langfuse-public-key"
7 os.environ["LANGFUSE_SECRET_KEY"] = "your-langfuse-secret-key"
8 
9 # Instrument BEFORE importing langfuse
10 LangfuseInstrumentor().instrument()
11 
12 # Now import and use langfuse normally
13 from langfuse.decorators import observe
14 from openai import OpenAI
15 
16 client = OpenAI()
17 
18 @observe()
19 def generate_joke():
20     response = client.chat.completions.create(
21         model="gpt-4o-mini",
22         messages=[{"role": "user", "content": "Tell me a joke"}],
23     )
24     return response.choices[0].message.content
25 
26 generate_joke()

How it works

The instrumentor patches Langfuse’s internal OTLP HTTP exporter to intercept spans and redirect them to Respan’s ingest endpoint. Existing Langfuse decorators (@observe, langfuse.trace()) continue to work unchanged — the data is simply mirrored to Respan.

Dependencies

langfuse
opentelemetry-api
opentelemetry-sdk
opentelemetry-instrumentation

Langfuse migration guide