Traceloop

Integrate Traceloop observability with PraisonAI agents

Traceloop Integration

Traceloop provides LLM observability with OpenTelemetry.

Setup

1. Install Dependencies

pip install traceloop-sdk

2. Set Environment Variables

export TRACELOOP_API_KEY=xxx

3. Initialize

from praisonai_tools.observability import obs

obs.init(provider="traceloop")

Usage

Auto-Instrumentation (Recommended)

from praisonai_tools.observability import obs
from praisonaiagents import Agent

obs.init(provider="traceloop")

# Everything is auto-traced!
agent = Agent(
    name="Assistant",
    instructions="You are a helpful assistant.",
    model="gpt-4o-mini",
)

response = agent.chat("Hello!")
print(response)

Explicit Tracing (For Fine Control)

from praisonai_tools.observability import obs
from praisonaiagents import Agent

obs.init(provider="traceloop", auto_instrument=False)

agent = Agent(
    instructions="You are a helpful assistant.",
    model="gpt-4o-mini",
)

with obs.trace("workflow"):
    response = agent.chat("Hello!")
    print(response)

Workflow Decorator

from praisonai_tools.observability.providers.traceloop_provider import TraceloopProvider

provider = TraceloopProvider()
provider.init()

@provider.workflow("my-workflow")
def my_workflow():
    # Your workflow code
    pass

Configuration Options

Option	Environment Variable	Description
api_key	TRACELOOP_API_KEY	Your Traceloop API key
app_name	-	Application name
disable_batch	-	Disable batching (default: False)