Usage & Logs

Cencori records every request that goes through the gateway. You see it in two places: charts on the AI Gateway overview, and individual requests in AI Gateway → Logs.

Open your project and go to AI Gateway → Overview. Pick a period at the top — Last hour, 24 hours, 7 days, 30 days, or All time — and every chart and table on the page updates together.

Usage

The Model usage, Cost by provider, Latency, and Errors charts show at most your plan's history: 7 days on Free, 30 days on Pro, 90 days on Team, and all time on Enterprise.

Requests

Below the charts, the five most recent requests in the period are listed, in the same format as Logs. Click a row for its full details, or View logs for the full list.

Go to AI Gateway → Logs to see every request. Each entry includes:

• Inputs and outputs: the full request and response
• Metadata: model, provider, latency, and status
• Usage and cost: tokens and cost, calculated from provider pricing

Filter by status, model, time range, and API key, or search. Click any request to open its details. See AI Request Logs for exporting and how far back you can view.

You can report your application's own web requests to Cencori:

import { Cencori } from 'cencori';
 
const cencori = new Cencori({ apiKey: process.env.CENCORI_API_KEY! });
 
// In your middleware or API route:
await cencori.telemetry.reportWebRequest({
  host: req.headers.get('host') || 'unknown',
  method: req.method,
  path: new URL(req.url).pathname,
  statusCode: response.status,
  latencyMs: Date.now() - startTime,
  userAgent: req.headers.get('user-agent') || undefined,
});

Or call the endpoint directly:

POST /api/v1/telemetry/web
Headers: { "CENCORI_API_KEY": "csk_..." }

The SDK method is fire-and-forget — it never throws and never blocks your application. Reported web requests are stored, but they are not currently shown in the dashboard.

To send usage to a FinOps tool, see Usage Export. To export request logs, see AI Request Logs.