Skip to main content
The OpenInference OpenAI instrumentor traces every call made with the OpenAI Python SDK — model, token usage, messages, and tool calls — as OTLP spans. Point the exporter at AgentMark and the traces arrive normalized.

​
Setup

1

Install the instrumentor and the OTLP exporter

pip install openinference-instrumentation-openai openai \
  opentelemetry-sdk opentelemetry-exporter-otlp-proto-http
2

Point the exporter at AgentMark and instrument OpenAI

Use your AgentMark API key and app id from project settings.
from openinference.instrumentation.openai import OpenAIInstrumentor
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.export import BatchSpanProcessor
from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter

provider = TracerProvider()
provider.add_span_processor(
    BatchSpanProcessor(
        OTLPSpanExporter(
            endpoint="https://api.agentmark.co/v1/traces",
            headers={
                "Authorization": "<YOUR_API_KEY>",  # raw key, no "Bearer" prefix
                "X-Agentmark-App-Id": "<YOUR_APP_ID>",
            },
        )
    )
)

OpenAIInstrumentor().instrument(tracer_provider=provider)
3

Run your calls

Use the OpenAI client as usual — client.chat.completions.create(...) or the Responses API. Each call arrives in AgentMark as a span, grouped into a trace. See Traces and logs.
Azure OpenAI is traced by the same instrumentor — use the AzureOpenAI client from the openai SDK and instrument it exactly as above.

​
What AgentMark captures

OpenAI spans use the OpenInference attribute conventions — model, token usage, input and output messages, tool calls, settings, and span kind are all mapped onto AgentMark’s normalized trace fields, and token counts feed cost tracking. See OpenInference for the full attribute mapping.

​
Next steps

OpenInference

How AgentMark reads OpenInference attributes

Traces and logs

Explore traces once they arrive

​
Have questions?

Reach out any time: