Customer identifier

You can pass customer identifier to your traces to get more insights into your LLM workflows.

You can only pass Keywords AI params to traces with Keywords AI tracing SDK. External integrations like OpenAI Agents SDK are not supported yet.

Example code

Example Python code

This is how to add Keywords AI params to your workflow traces:

It’s the same way to add Keywords AI params to a workflow or a task.

agent.py

@workflow(name="my_workflow")
def my_workflow():
    # Add Keywords AI params to the current trace
    with keywordsai_span_attributes(
        keywordsai_params={
            "customer_params": {
                "customer_identifier": "123",
            },
            "metadata": {"some_key": "some_value"},
        }
    ):
        # Your LLM calls or other operations here
        client = OpenAI()
        response = client.chat.completions.create(
            model="gpt-4o",
            messages=[{"role": "user", "content": "Hello, world!"}],
        )
    return response

Example JS/TS code

This is how to add Keywords AI params to your workflow traces:

It’s the same way to add Keywords AI params to a workflow or a task.

Once you integrate the Keywords AI tracing SDK in your project, and annotate the workflows and tasks, you can easily add Keywords AI params in your traces.

Here’s an example of how to add Keywords AI params to a workflow:

agent.tsx

async function testWorkflow() {
  return await keywordsAI.withWorkflow({ name: "test" }, 
  async () => {
    keywordsAI.withKeywordsAISpanAttributes(
      async () => {
        testTask();
      },
      {
        customer_params: {
          customer_identifier: "123",
        },
        metadata: { some_key: "some_value" },
      }
    );
    return "test workflow";
  });
}