How to Track API Usage and Manage Extract/Crawl Endpoints on the Tavily API Platform

The Tavily API Platform (app.tavily.com) is the console where you generate and monitor API keys, track usage, and manage the Extract and Crawl endpoints — all in one place. If you need to know how much of your quota you've consumed or where to configure the Extract/Crawl capabilities your application calls, that work happens in this dashboard rather than in your code. The steps below describe what each area is for and how to use it; exact menu labels and metric names may differ slightly from what you see after signing in.

What the platform covers

According to the platform's own description, the console is organized around three jobs:

• API keys — generate new keys and monitor their activity.
• Usage — track how much you're calling the API.
• Extract / Crawl endpoints — manage the endpoints that power extraction and crawling.

That grouping matters because usage tracking and endpoint management are related: the endpoints you enable determine what shows up in your usage data, and your usage data tells you whether those endpoints are behaving as expected.

Tracking API usage

Where to look

Usage tracking lives in the platform dashboard, not in a separate billing portal. After signing in, the usage view is the place to see consumption tied to your account and keys.

What to watch

When reviewing usage, focus on these dimensions:

How to use it

1. Establish a baseline. Check usage before a change (a new deployment, a batch job, a crawler run) so you have something to compare against.
2. Attribute spikes. If volume jumps, use the per-key view to identify which key is responsible. A single key spiking usually means one service or one runaway job.
3. Correlate with endpoint activity. If Extract or Crawl calls dominate, the issue is likely in how you're invoking those endpoints rather than in general API traffic.

Managing Extract and Crawl endpoints

What "managing" means here

The platform lets you manage the Extract and Crawl endpoints from the console. In practice this means controlling which of these capabilities are available to your keys and integrations, and reviewing their activity alongside your other usage.

A practical workflow

1. Confirm which endpoints you actually need. Extract and Crawl serve different purposes — extraction pulls structured content from a known page, while crawling discovers and traverses pages. Enabling only what you use keeps usage data readable.
2. Tie endpoints to specific keys. If different applications use Extract versus Crawl, give them separate keys so usage attribution stays clean.
3. Review endpoint activity after changes. Any time you adjust how an endpoint is called, recheck the usage view to confirm the change had the intended effect.

Using this to optimize calls or debug problems

The console is most useful when you treat usage data as a diagnostic signal:

• Unexpected volume on Crawl often points to a crawl that's following more links than intended, or a job that's re-running.
• Extract calls that don't match your request count can indicate retries in your client code.
• A key with activity you can't explain is worth rotating or disabling until you identify the source.

Because the platform combines key management, usage monitoring, and endpoint management in one place, you can move from "something looks wrong" to "here's the key and endpoint responsible" without switching tools.

What to verify yourself

The platform description confirms the three capabilities (keys, usage, Extract/Crawl management) but does not specify exact metric names, retention windows, or plan limits. When you sign in, check the usage view directly for the metrics it exposes, and confirm any quota or pricing details on the platform itself rather than assuming them.