Track your first request, then add the fields that make the analysis smarter

Quickstart

Initialize ZenLLM once, configure your team or customer context, and use the provider wrapper in place of your direct OpenAI client.

import zenllm
from zenllm import OpenAI

zenllm.init(api_key="zllm_your_api_key")
client = zenllm.get_client()
client.configure(team_id="engineering", customer_id="acme-corp")

llm = OpenAI(api_key="sk-your-openai-key")
response = llm.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Summarize this document"}],
    workflow_type="summarization",
)

Context telemetry

If your workload is conversational, agentic, or retrieval-heavy, send the new explicit fields so ZenLLM can separate history growth from static prompts and retrieval overhead.