Site Looker Atter capture + extract API

Capture the page. Keep the content.

Capture any website as an image or PDF, or extract its readable content as Markdown, JSON, plain text or HTML. Paste a URL below.

Output
Format
Device

Examples

Tap a card to prefill the form.

API reference

One GET request. No key, no signup. Returns the image bytes.

GET https://api.sitelookeratter.com/screenshot?url=https://example.com&device=mobile&removeCookies=true
url
Page to capture. Required. https:// is added if you omit it.
device
mobile · tablet · laptop · desktop · iphone · android - preset viewport + user agent.
width, height
Custom viewport in px. Used when device is omitted.
format
png · jpeg · webp · pdf. Default png. pdf renders the whole page as A4.
quality
1–100. Applies to JPEG and WebP only. Default 80.
scale
1–3 device pixel ratio. 2 = crisp retina captures. Default 1.
delay
Milliseconds to wait before capture. 0–10000. Use waitFor instead when you can name the element.
waitFor
CSS selector to wait for before capturing, up to 15 seconds. The reliable way to capture a page that builds itself in the browser. Returns 400 if it never appears, rather than capturing the wrong moment.
full
true captures the whole page height, not just the viewport.
dark
true forces the page into its dark theme.
removeCookies
true strips cookie-consent banners.
blockAds
true blocks ad slots.
removeChat
true hides chat / support widgets.
removePopups
true dismisses newsletter and promo popups.
selector
CSS selector - capture a single element.
hide
Comma-separated CSS selectors to hide before capture.
useragent
Custom user agent string. Used when device is omitted.

Content extraction

The same page as clean, readable content - ideal for feeding to a language model. Cached 24 hours, same no-key access. 

GET https://api.sitelookeratter.com/extract?url=https://example.com/article
url
Page to read. Required. https:// is added if you omit it.
format
markdown · json · text · html. Default markdown. json returns title, byline, excerpt, word count and the content in all three forms.

MCP server

Both tools in any client that speaks the Model Context Protocol. Nothing to configure and no key to obtain. 

https://api.sitelookeratter.com/mcp
Claude Code
claude mcp add --transport http sitelookeratter https://api.sitelookeratter.com/mcp
Other clients
Add the URL above as a remote (streamable HTTP) server. It is stateless and needs no auth header.
Tools
screenshot and extract, the same two endpoints with the same parameters.
PDF
MCP carries images but not PDFs, so use the HTTP endpoint above for format=pdf.

Access

Free, with no key and no account. 50 requests per hour per IP address, which is there to keep the render queue available to everyone rather than to sell you a bigger number.

Rate limit
Every response carries X-RateLimit-Limit, -Remaining and -Reset. Over the limit is a 429 with Retry-After.
Cache
24 hours, keyed on the whole parameter set. A cached response sends X-Cache: HIT and skips the render queue, but it still counts against the rate limit.

Frequently asked questions

How do I take a screenshot?

Paste a website URL, pick your options (device, format, content blockers) and hit Generate. The API captures the page and returns a clean image instantly.

What formats are supported?

Capture pages as PNG, JPEG, WebP or PDF. Extract their readable content as Markdown, JSON, plain text or HTML.

What does content extraction return?

It renders the page like a browser, finds the main readable content and removes the surrounding navigation and clutter. Choose Markdown or plain text for writing and language models, HTML for reuse, or JSON for all formats plus metadata.

Can I capture full-page screenshots?

Yes. Turn on Full Page to capture the entire page height, not just the visible viewport - ideal for long articles and landing pages.

Does it work with international websites?

Absolutely. Emoji and every major script render correctly - Chinese, Japanese, Korean, Arabic, Hindi, Thai and more, thanks to full Unicode support.

Can I use it from Claude or another AI assistant?

Yes. There is an MCP server at https://api.sitelookeratter.com/mcp. Add that URL to any client that supports remote MCP servers and the screenshot and extract tools appear with nothing to configure and no key to obtain. In Claude Code it is one command: claude mcp add --transport http sitelookeratter https://api.sitelookeratter.com/mcp

Are there any limits?

Fifty requests per hour per IP address, on both capture and extraction. Every response tells you where you stand in the X-RateLimit headers, and going over returns a 429 with Retry-After. The limit exists to keep the render queue available to everyone, not to sell you a bigger one.

Is this free to use?

Yes. The service is free while in beta. We are gathering feedback to build the best screenshot API possible.

What makes this different?

It strips cookie banners, ads and chat widgets for you, and it is built on Bun for fast captures.