Screenshots

/screenshot opens a URL in a headless Chrome browser, captures it, stores the image and returns a public URL to it. Cookie banners and ad slots are hidden on every capture. You choose:

What to capture: the viewport, the full scrollable page (full_page) or one element (selector).
What to render it as: a viewport size (width, height), or a real device (device) with its viewport, pixel density, touch support and user agent, so the page serves its phone or tablet layout.
How the page looks: its dark theme (dark_mode), and with ads and trackers blocked (block_ads).
When to capture: after a delay (wait_for), or once an element appears (wait_for_selector).
The output: jpeg (smaller), png (lossless), webp (best compression) or pdf (the whole page as a document), with quality for JPEG and WebP.

Example requests

curl "https://api.prefetch.io/screenshot?url=https://stripe.com" \
  -H "X-API-Key: $PREFETCH_API_KEY"

curl "https://api.prefetch.io/screenshot?url=https://github.com&device=iphone_15&dark_mode=true" \
  -H "X-API-Key: $PREFETCH_API_KEY"

curl "https://api.prefetch.io/screenshot?url=https://stripe.com/pricing&selector=%23pricing&format=png" \
  -H "X-API-Key: $PREFETCH_API_KEY"

curl "https://api.prefetch.io/screenshot?url=https://www.theverge.com&full_page=true&block_ads=true&format=webp&quality=70" \
  -H "X-API-Key: $PREFETCH_API_KEY"

const params = new URLSearchParams({
  url: "https://stripe.com",
  device: "pixel_7",
  format: "webp",
  block_ads: "true",
});

const res = await fetch(`https://api.prefetch.io/screenshot?${params}`, {
  headers: { "X-API-Key": process.env.PREFETCH_API_KEY },
});
const { data } = await res.json();
console.log(data.screenshot_url);

import os, requests

r = requests.get(
    "https://api.prefetch.io/screenshot",
    params={"url": "https://stripe.com", "device": "pixel_7", "format": "webp", "block_ads": "true"},
    headers={"X-API-Key": os.environ["PREFETCH_API_KEY"]},
)
print(r.json()["data"]["screenshot_url"])

Example response

{
  "success": true,
  "data": {
    "url": "https://github.com",
    "screenshot_url": "https://assets.prefetch.io/screenshots/github-iphone15.jpg",
    "viewport": { "width": 393, "height": 659 },
    "format": "jpeg",
    "full_page": false,
    "captured_at": "2026-10-13T09:36:12.000Z",
    "device": "iphone_15",
    "device_scale_factor": 3,
    "dark_mode": true,
    "selector": null
  },
  "meta": {
    "requestId": "c1d2e3f4-5a6b-7c8d-9e0f-1a2b3c4d5e6f",
    "durationMs": 2450
  }
}

viewport is in CSS pixels. The image is viewport × device_scale_factor pixels, so an iphone_15 capture is 1,179 pixels wide. warnings appears when something wasn’t as you asked but the capture still happened, for example a wait_for_selector that never appeared.

Devices

`device`	Viewport	Density
`iphone_15`	393 × 659	3×
`iphone_15_pro_max`	430 × 739	3×
`iphone_se`	320 × 568	2×
`pixel_7`	412 × 839	2.625×
`galaxy_s8`	360 × 740	3×
`ipad`	810 × 1080	2×
`ipad_mini`	768 × 1024	2×
`ipad_pro_11`	834 × 1194	2×
`galaxy_tab_s4`	712 × 1138	2.25×
`desktop_retina`	1280 × 720	2×

A device sets the viewport, so width and height are ignored when you pass one.

Elements

selector captures the first element that matches a CSS selector, whole: #pricing, header, .hero img. Prefetch waits up to 10 seconds for it. If nothing matches, the request fails with 422 and TARGET_ELEMENT_NOT_FOUND, and isn’t charged. An element can’t be captured as a PDF. wait_for_selector is different: it’s best effort. The page is captured once the selector appears, or after 10 seconds, with a warning, if it never does.

Formats

Format	Best for	Notes
`jpeg`	General use, thumbnails	Smallest files, lossy. `quality` 1–100, default 85.
`png`	Logos, UI elements, exact pixels	Lossless, larger files.
`webp`	Web display	Best size for the quality. `quality` 1–100, default 85.
`pdf`	Documents, archiving	The whole page, printed with backgrounds.

Images larger than 8 MB, such as a long page at a phone’s pixel density, are scaled down until they fit.

Limits

Parameter	Default	Range
`width`	1440	320–2560
`height`	900	240–5000
`format`	`jpeg`	`jpeg`, `png`, `webp`, `pdf`
`full_page`	`false`
`device`	none	See Devices
`dark_mode`	`false`
`selector`	none	A CSS selector, up to 200 characters
`wait_for`	0	0–30,000 ms
`wait_for_selector`	none	A CSS selector, up to 200 characters
`block_ads`	`false`
`quality`	85	1–100

Credits

1 credit per capture, whatever the options. Failures are never charged.

Batch screenshots

Capture up to 500 URLs in one job with POST /batch/screenshot.

API reference

Every parameter, and a playground to try it.

Get started

Agents and MCP

Company data

Web data

Jobs and monitoring

Platform

Going to production

Example requests

Example response

Devices

Elements

Formats

Limits

Credits

Next

Batch screenshots

API reference

​Example requests

​Example response

​Devices

​Elements

​Formats

​Limits

​Credits

​Next

Batch screenshots

API reference

Example requests

Example response

Devices

Elements

Formats

Limits

Credits

Next