A tracker turns a web page into structured data that refreshes on a schedule. Describe what to collect and Equibles builds a fixed collection plan, tests it, and only then starts collecting. Trackers are a Pro feature; [Pricing](/pricing) shows how many Pro includes.

## Create a tracker

1. Open [Trackers](/trackers).
2. Under **What data should Equibles collect?**, give the exact page URLs, the fields you want, and any read-only navigation needed to reach them.
3. Pick a schedule and click **Generate and test**.

A tracker is private unless you tick **Public** under **Visibility**, which lets Equibles show its latest verified result to anyone. A tracker that signs in to a site always stays private.

- Generating a tracker spends 3 [ALVIS credits](/docs/alvis#credits). Scheduled collections spend none.
- If the request needs more detail, you are told what is missing and the credits are refunded.
- Recurring collection starts only after a test returns a verified result.
- The tracker gets a generated name, which you can change later.

## Schedule

Choose every hour, every 4, 8 or 12 hours, every day, every week or every month (every 30 days). Change it later with the tracker's **Schedule** dropdown and **Save**, or **Pause** and **Resume** it.

## Results and history

**Show data** or **History and connection details** opens the tracker's page: its latest verified collection, a timeline of past captures and the collection history.

- A failed collection never erases the last good result.
- History keeps up to 12 months, 4,000 observations or 20 MB, whichever comes first; the oldest entries go first.
- **Improve extraction** refines what the tracker collects. It spends 3 credits, which are not refunded if the live test fails; your previous instructions and data stay in place.
- **Retry testing**, **Retry collection** and **Repair extraction** recover a tracker that failed, and a paused tracker offers **Retry and resume** or **Repair and resume**.
- **Make public** or **Make private** changes visibility, and **Delete** removes the tracker with all its past data.

## Sites you sign in to

For pages behind a login, add the account under **Site connections** first. Each connection is limited to one website, and you can save up to 5.

## Read trackers from your AI or code

Your trackers' latest and past results are readable through the [tracker MCP tools](/docs/mcp/tools/web-data) and the [tracker REST endpoints](/docs/api/endpoints/web-data-feeds). Both read stored results and never contact the source website.

For step-by-step help in the app, see the [trackers guide](/trackers/guide).