# Full-page screenshots Capture a page from top to bottom, load lazy images on the way, and keep very tall pages under control. Add `full_page=true` and the capture runs to the bottom of the page. ```bash curl "https://curlshot.com/api/v1/screenshot?access_key=YOUR_ACCESS_KEY&url=https://en.wikipedia.org/wiki/Eiffel_Tower&full_page=true" \ --output eiffel-full.png ``` You get one tall image with the whole article in it. ![The complete Wikipedia article about the Eiffel Tower in one tall image](https://curlshot.com/docs/examples/wikipedia-full-page.webp) Without the option, you get only the first screen. That is the part a visitor sees before scrolling. ```bash curl "https://curlshot.com/api/v1/screenshot?access_key=YOUR_ACCESS_KEY&url=https://en.wikipedia.org/wiki/Eiffel_Tower&viewport_height=800" \ --output eiffel-first-screen.png ``` ![The top of the Wikipedia article about the Eiffel Tower](https://curlshot.com/docs/examples/wikipedia-viewport.webp) ## How it works The width of the image still comes from the viewport. The viewport is the browser window the page is drawn in. The height is as tall as the page turns out to be. So [`viewport_width`](https://curlshot.com/docs/options.md#viewport_width) and [`viewport_device`](https://curlshot.com/docs/devices.md) still matter. A full-page capture on a phone preset gives you the long mobile version of the page. ### full_page Capture the whole page, not only the first screen. The default is `false`. ## Lazy images Many sites load images only when you scroll near them. This is called lazy loading. A capture taken without scrolling would show empty boxes further down the page. To avoid that, a full-page capture first scrolls through the page from top to bottom. It gives the images a moment to arrive, then goes back to the top and takes the screenshot. ### full_page_scroll Scroll through the page before the capture so lazy-loaded images appear. The default is `true`. You rarely need to change it. Turn it off in two cases: - The page is short and has no lazy images, and you want the fastest render. - Scrolling changes the page in a way you do not want. Some sites load more and more content as you scroll, or shrink their header. ```bash curl "https://curlshot.com/api/v1/screenshot?access_key=YOUR_ACCESS_KEY&url=https://en.wikipedia.org/wiki/Eiffel_Tower&full_page=true&full_page_scroll=false" \ --output eiffel-no-scroll.png ``` > **Note** > > The scroll has a time budget. On a very long page with slow images, a few may still be missing. Adding a [`delay`](https://curlshot.com/docs/options.md#delay) does not extend the scroll. A lower `full_page_max_height` does help, because there is less to scroll through. ## Very tall pages Some pages never end. News feeds and shop listings keep adding content. To keep the image a usable size, a full-page capture stops at a maximum height. ### full_page_max_height Cut a full-page capture off at this height, in pixels. The default is `20000`. The smallest value is `100` and the largest is `30000`. A page shorter than the limit is not stretched. A page taller than the limit is cut off at the limit, and you still get a valid image. This request keeps only the first 5000 pixels of the article: ```bash curl "https://curlshot.com/api/v1/screenshot?access_key=YOUR_ACCESS_KEY&url=https://en.wikipedia.org/wiki/Eiffel_Tower&full_page=true&full_page_max_height=5000" \ --output eiffel-5000.png ``` The height is counted in page pixels. With a pixel density above 1, the file has more pixels than that. A 5000 pixel page at [`device_scale_factor=2`](https://curlshot.com/docs/options.md#device_scale_factor) is 10000 pixels tall in the file. Very large captures can hit other limits too: - At a high pixel density, the capture may be cut shorter than `full_page_max_height` to keep the total pixel count within bounds. - If the finished file is too big, you get [`content_too_large`](https://curlshot.com/docs/errors.md#content_too_large). Use `format=jpeg` or `format=webp`, a lower [`image_quality`](https://curlshot.com/docs/options.md#image_quality), or a smaller `full_page_max_height`. ### Limits of your plan A full-page screenshot counts as one screenshot against your quota, however tall it is. What a plan limits is the size: - **Height.** A plan can have a lower maximum height than `30000`. A larger `full_page_max_height` is lowered to the plan's maximum, and the capture is cut off there. - **Pixel density.** A plan can limit the [`device_scale_factor`](https://curlshot.com/docs/options.md#device_scale_factor) of a full-page capture. A higher value is lowered to the plan's maximum. This also applies to the density a [device preset](https://curlshot.com/docs/devices.md) brings along. In both cases you still get a valid screenshot, not an error. Screenshots of the first screen and [element captures](https://curlshot.com/docs/element.md) keep the density you asked for. Your own values are `full_page_max_height` and `full_page_max_scale` in the answer of [`GET /usage`](https://curlshot.com/docs/usage-and-limits.md). > **Tip** > > For tall pages, JPEG or WebP is usually the better choice. A PNG of a 20000 pixel page can be many megabytes. ## Sticky headers and floating bars A sticky header is a menu bar that stays at the top of the window while you scroll. Cookie bars and chat bubbles float in a similar way. In a full-page capture these can land in odd places. A bar may cover part of the content, or appear in a spot where it makes no sense. The fix is to hide the element before the capture with [`hide_selectors`](https://curlshot.com/docs/options.md#hide_selectors). A selector is a short pattern that points at an element, such as `.site-header` for an element with the class `site-header`. ```bash curl "https://curlshot.com/api/v1/screenshot?access_key=YOUR_ACCESS_KEY&url=https://en.wikipedia.org/wiki/Eiffel_Tower&full_page=true&hide_selectors=.vector-sticky-header" \ --output eiffel-clean.png ``` To find the right selector, open the page in your browser, right-click the bar and choose **Inspect**. Look for its `class` or `id`. For consent pop-ups and chat widgets, try [`block_cookie_banners`](https://curlshot.com/docs/blocking.md) and `block_chats` first. They cover the common ones without a selector. > **Common mistakes** > > - **Combining it with an element capture.** When you send [`selector`](https://curlshot.com/docs/element.md) or the `clip_*` options, those decide the area and `full_page` is not used. > - **A page that looks cut off.** It probably reached `full_page_max_height`. Raise it, up to `30000` or the maximum of [your plan](#limits-of-your-plan). > - **A full-page phone screenshot that is less sharp than expected.** Your plan limits the pixel density of full-page captures. See [Limits of your plan](#limits-of-your-plan). > - **Blank areas in the middle.** The content there appears only after an animation or a script. Try [`wait_until=networkidle`](https://curlshot.com/docs/waiting.md) or a short `delay`. > - **Timeouts on heavy pages.** Long pages take longer. Raise [`timeout`](https://curlshot.com/docs/options.md#timeout), or run the request with [`async=true`](https://curlshot.com/docs/async-and-webhooks.md). ## Where to go next - [Capturing an element](https://curlshot.com/docs/element.md): when you need one part of the page, not all of it. - [PDF rendering](https://curlshot.com/docs/pdf.md): a long page as a document with real pages. - [Waiting and timing](https://curlshot.com/docs/waiting.md): make sure the page is ready before the capture. - [Blocking ads, cookie banners, trackers and chats](https://curlshot.com/docs/blocking.md): remove the usual floating clutter.