> ## Documentation Index
> Fetch the complete documentation index at: https://docs.pagepith.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Scrape web pages

> Control caching, timeouts, and dynamic page extraction.

Use the synchronous scrape endpoint when your caller can wait for the result. PagePith tries the least expensive extraction method first and escalates when a page requires JavaScript rendering or additional handling.

## Request options

| Field             | Default  | Use it when                                                   |
| ----------------- | -------- | ------------------------------------------------------------- |
| `url`             | Required | You have a complete public `http` or `https` URL.             |
| `forceFresh`      | `false`  | You need to bypass a cached result.                           |
| `timeoutMs`       | `45000`  | You need a request budget from 1,000 to 120,000 milliseconds. |
| `waitForSelector` | None     | Dynamic content is ready only after a CSS selector appears.   |
| `extraWaitMs`     | `0`      | A rendered page needs up to 10,000 milliseconds to settle.    |

```bash theme={null}
curl --request POST \
  --url https://api.pagepith.com/v1/api/scrape \
  --header "Authorization: Bearer $PAGEPITH_API_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "url": "https://example.com/pricing",
    "waitForSelector": "main",
    "extraWaitMs": 1000
  }'
```

## Cache behavior

Responses include a `cached` boolean. Set `forceFresh` to `true` when freshness matters more than latency. You can also invalidate a URL explicitly before its next scrape.

<Tip>
  Use cached results for repeat reads. Reserve `forceFresh` for content you expect to have changed.
</Tip>

## Choose synchronous or asynchronous scraping

* Use [synchronous scraping](/api-reference/scrape/scrape-url) for interactive requests and single URLs.
* Use [asynchronous scraping](/guides/async-scraping) for batch ingestion or work that should not hold an HTTP connection open.