Monitor your OpenAI Agents SDK agents in production.

Agents built with the OpenAI Agents SDK run tools, hand off between agents and call models in a loop. Agent Ping captures one record per run so you can see what each cost, whether it ran, and whether the output held up.

Track your first AI workflow Try it live first, no signup Read the OpenAI Agents SDK docs

Production AI in OpenAI Agents SDK fails quietly.

An OpenAI Agents SDK agent can hand off, retry and call a model several times in a single run. The provider console shows tokens by API key, not which agent or run spent them, and nothing tells you when a scheduled agent simply stopped firing.

Wire it up in a few lines.

Start a run, attach the Agent Ping hooks to your Runner, and hand the Agents SDK an instrumented OpenAI client. Model calls, tool calls and handoffs land on one run with cost, status and latency. In Python, agentping.instrument_openai_agents() does all of this in one line.

Install

install OpenAI Agents SDK 
npm install @agentping/sdk @openai/agents openai

Instrument

example · typescript OpenAI Agents SDK 
import { Agent, Runner, setDefaultOpenAIClient } from "@openai/agents";
import OpenAI from "openai";
import * as agentping from "@agentping/sdk";

agentping.init({ apiKey: process.env.AGENTPING_API_KEY });

const run = agentping.run("support-triage", { customerId: "acme-corp" });

// Model calls: give the Agents SDK an instrumented OpenAI client.
setDefaultOpenAIClient(agentping.instrumentOpenAI(new OpenAI(), { run }));

// Tool calls, handoffs and agent turns: attach the hooks to the runner.
const runner = new Runner();
new agentping.AgentPingHooks(run).attach(runner);

const result = await runner.run(triage, "My invoice is wrong");
await run.finish({ status: "success" });

Full setup, environment variables and options are in the OpenAI Agents SDK docs.

What Agent Ping tracks for OpenAI Agents SDK.

  • Cost per agent run, attributed to the customer and feature you tag.
  • Tool calls and agent handoffs captured as a single run trace.
  • Run status, latency and retry count, so a stuck run is visible.
  • Missed run alerts for agents on a schedule.
  • Optional output quality scoring on a sampled subset of runs.
Read the OpenAI Agents SDK docs Heartbeat webhook docs

What you see in the dashboard.

  • A run list for OpenAI Agents SDK, each row showing cost, status and latency, filterable by agent, customer and feature.
  • Cost by agent, customer and feature, so you can see which accounts and features drive AI spend, not just one provider total.
  • An incident the moment a scheduled run misses its window or a cost spike breaks its baseline, delivered to Slack, email, PagerDuty or a webhook.
  • A quality trend once you turn on scoring, so a regression in your OpenAI Agents SDK output shows up as a falling line, not a support ticket.
How do I monitor an agent built with the OpenAI Agents SDK?
Install the Agent Ping SDK, start a run, and attach AgentPingHooks to the agent run. In Python, call agentping.instrument_openai_agents() once and every Runner.run carries the hooks. In TypeScript, attach new AgentPingHooks(run) to your Runner and give the Agents SDK an instrumented OpenAI client, as in the example above. Each run then lands as one Agent Ping record with a model call per request, a tool call per tool and a step per handoff, priced server side, with alerts for failures, missed scheduled runs and cost spikes.
Does this work with agent handoffs?
Yes. A handoff between agents is captured within the run trace, so a multi-agent run is one record with the tool calls and child runs attached, and the cost is attributed to the run that spent it.
Will it slow my agent down?
No. The SDK never blocks your agent. Telemetry sends on a background thread with a hard 2-second timeout, and if Agent Ping is unreachable your agent runs as if nothing is installed.
What does it cost to monitor OpenAI Agents SDK?
There is a free plan with no card and no time limit, enough to wire up your first agents. Paid plans start at $99/month with named limits and no metered billing, so the bill never surprises you. Annual billing is two months free.
Does Agent Ping see my prompts and outputs?
Only if you choose to send them. By default Agent Ping prices runs from model and token counts and tracks status and timing, which needs no prompt or output content. Payload capture for quality scoring is opt-in and configurable per agent.

Monitor your OpenAI Agents SDK AI workflows.

Send one run from OpenAI Agents SDK and see cost, status and latency. Then add an alert for cost spikes or missed runs.

Track your first AI workflow Try it live first, no signup Read the OpenAI Agents SDK docs

Step-by-step guide: monitor OpenAI Agents SDK agents → Spend, cost attribution → Pulse, live monitoring → Verify, quality scoring →