Auto-Instrumentation
SDKAutomatic capture for OpenAI, Anthropic, LangChain, and LiteLLM.
Usage Example
example_usage.py
# Example integration for Auto-Instrumentation
from agentis import init, observe
# Initialize Agentis with auto-instrumentation
init(
project="production-fleet",
environment="production",
redact_pii=True
)
@observe(name="auto-instrumentation")
def execute_task(payload: dict):
# Traces, spans and costs are tracked with 0ms added latency
return {"status": "success", "data": payload}
Key Capabilities
- OpenTelemetry OTLP standard semantics across all parent and child spans.
- Exact span-end token pricing computed locally across 2,500+ foundation models.
- Client-side PII scrubbing before telemetry packets leave your application memory.