dr-langfuse

Typed Pydantic models, adapter protocols, and concrete implementations for Langfuse prompt fetching and trace emission.

Install

pip install dr-langfuse

Quick start

from dr_langfuse import LangfusePromptProvider, PromptFetchRequest

provider = LangfusePromptProvider()  # reads LANGFUSE_* env vars
payload = provider.fetch_prompt(
    PromptFetchRequest(prompt_name="my-prompt", variables={"topic": "recursion"})
)
print(payload.system_content)
print(payload.task_content)

from dr_langfuse import LangfuseTraceEmitter, TraceEventRequest

emitter = LangfuseTraceEmitter()
ack = emitter.emit_trace(
    TraceEventRequest(event_name="eval.step", tags=["experiment"], metadata={"run": 1})
)
print(ack.accepted, ack.trace_id)

What's included

All symbols are re-exported from the top-level package.

Configuration

Set standard Langfuse environment variables:

export LANGFUSE_PUBLIC_KEY="pk-..."
export LANGFUSE_SECRET_KEY="sk-..."
export LANGFUSE_BASE_URL="https://cloud.langfuse.com"  # optional
# LANGFUSE_HOST is still supported as a deprecated fallback.

Or pass config explicitly:

from dr_langfuse import LangfuseConfig, LangfusePromptProvider

config = LangfuseConfig(public_key="pk-...", secret_key="sk-...", host="https://cloud.langfuse.com")
provider = LangfusePromptProvider(config=config)

Requirements

• Python 3.12+
• pydantic 2.x
• langfuse 3.x

License

MIT