Control plane for AI agents

Monitor traces, costs, and failures across every model provider — one SDK, full visibility in production.

Try for free →See the dashboard

Dashboard

24H7D1M

Traces

12,847

last 30 days

Cost

$142.50

estimated

Latency

1.2s

p50

Success

99.2%

rate

Requests over time

Top models

gpt-4o-mini6,241
claude-3-5-haiku3,102
gemini-2.0-flash1,890

Observability built for AI agents

Connect your agents and every metric flows from real telemetry — not demo data.

Trace Explorer

Search and inspect execution traces with event timelines for every agent run.

Cost & Usage

Token usage, cost by model, latency percentiles, and anomaly detection.

Agents

Agents auto-discover when you instrument with the SDK — health and metadata.

Trace Monitoring

Success rates, latency, and failure trends per agent across your fleet.

Alerts

Threshold monitoring for failure rate, latency, cost, and errors.

One SDK for 13+ providers

Instrument any major AI provider with Traceplane. Switch models without rewriting your telemetry.

import { init, trace } from "traceplane";
import OpenAI from "openai";

init({
  apiKey: process.env.TRACEPLANE_API_KEY!,
  baseUrl: process.env.TRACEPLANE_API_URL ?? "/api",
});

await trace(
  { agent: "sdk-agent", model: "gpt-4o-mini", provider: "openai" },
  async (run) => {
    const prompt: string = "Hello!";
    run.setInput(prompt);
    const client = new OpenAI({ apiKey: process.env.OPENAI_API_KEY! });
    const response = await client.chat.completions.create({
      model: "gpt-4o-mini",
      messages: [{ role: "user", content: prompt }],
    });
    return response.choices[0]?.message?.content ?? "";
  },
);