How to Screenshot Every Page of a Website (Build a Visual Site Audit)

You are redesigning a site, migrating a CMS, or auditing a competitor, and you need a screenshot of every page, not just the one you happen to have open. The head query for this reads two ways. One is manual: capture the single long page in front of you. The other is the job most developers actually mean here: turn a whole site into one image per page for a visual inventory. This guide is about the second one.

The short answer

To screenshot every page of a website, do two things in order:

1. Enumerate the routes. Get a flat list of the site's URLs, usually from its sitemap.xml or a small crawl step you run.
2. Loop the list through a capture. Send each URL to a screenshot API one at a time and collect the hosted image for every page.

There is no magic "shoot the whole site in one call" primitive. Every tool does the same thing underneath: it visits each page in turn and renders it. The real work is building the route list and driving the loop.

Why the manual tools stop at one page

The top of this SERP is full of single-page capture tools: the Chrome command menu's "Capture full size screenshot," Edge's Web capture, and extensions like GoFullPage. They are excellent at one thing, capturing the entire scrollable length of the page you are currently on. None of them screenshot a site. You would have to open every page yourself and press the capture key on each, which is exactly the tedium you are trying to avoid on a fifty-page audit.

For a true whole-site job you need the route list and a loop, which means code.

Step 1: enumerate the site's routes

Before you can capture every page, you need to know every page. Two reliable sources:

• The sitemap. Most sites publish https://example.com/sitemap.xml, an XML file listing their canonical URLs. Fetch it, parse the <loc> entries, and you have your list. This is the cleanest source because the site owner already curated it.
• A crawl step. No sitemap? Run a small crawler that starts at the homepage, follows internal links, and collects unique routes. Keep it scoped to the site's own domain so you do not wander off.

Either way you end up with a plain array of URLs. That array is the only input the capture loop needs.

One honest note up front: a screenshot API captures the URLs you hand it. It does not crawl your site or follow links on its own. The enumeration in this step is a job you run; the API's job starts once you have the list.

Step 2: loop the routes through a screenshot API

With the list in hand, each capture is a single HTTP request. You send a URL, you get back a hosted image. No browser to install, nothing to keep patched. Here is the full pattern in Node: read the routes, capture each one full-page, and collect the results.

// routes.json is your enumerated list, e.g. parsed from sitemap.xml
import { readFileSync } from 'node:fs';

const routes = JSON.parse(readFileSync('routes.json', 'utf8'));
const audit = [];

for (const url of routes) {
  const res = await fetch('https://api.grabbit.live/v1/grabs', {
    method: 'POST',
    headers: {
      Authorization: 'Bearer sk_live_...',
      'Content-Type': 'application/json',
    },
    body: JSON.stringify({ url, width: 1280, full_page: true, format: 'webp' }),
  });
  const data = await res.json();
  audit.push({ url, image_url: data.image_url });
}

console.table(audit);

Each response includes a hosted image_url, so you can drop the results straight into a CSV, a database, or a static gallery page. Because every request is independent, one page that 404s or hangs never sinks the rest of the audit. This is the same repeatable, scriptable job Grabbit's automated screenshots are built for, and you can wire it up with a free test key before adding a card. For the mechanics of driving a flat list of URLs, see how to screenshot a list of URLs; a whole-site audit is that same loop, fed by the routes you enumerated in step 1.

The options that matter for a whole-site audit

Each call in the loop is configured on its own, so a mixed set of pages is no problem:

• full_page: true captures the entire scrollable page, not just the viewport. Essential for an audit where you want the whole page on record.
• width sets the viewport (320 to 1920). Use one consistent width across the whole site so the pages line up when you compare them side by side.
• format can be png, jpeg, or webp. WebP keeps a large audit small on disk; PNG is worth it when you want lossless baselines.
• delay_ms waits up to 10 seconds before the capture, which helps on pages that fade content in after load.
• selector captures just one element by CSS selector, useful if your audit only cares about, say, the header or a pricing table on every page.

For a large site where you do not want to hold a connection open per request, the same endpoint accepts Prefer: respond-async (or ?async=true) and returns immediately with a job you poll later. That keeps a several-hundred-page run from blocking your script.

Turn the captures into an audit

The reason to screenshot every page is rarely the images themselves, it is what you do with them:

• Redesign inventory. Before a rebuild, capture the current site so you have a visual record of every page you are replacing. Nothing gets forgotten in the migration.
• Visual QA after a deploy. Re-run the same route list on staging and production and eyeball the pages that changed.
• Competitor teardown. Snapshot a competitor's key pages, then re-run the list next month. Save each image_url next to its source URL and a timestamp and the diff over time tells a story, which is the foundation of website change monitoring.

Because the loop is deterministic and hands-off, you can run it on a schedule with cron or a CI job and have a fresh visual snapshot of the whole site whenever you need one.

Wrapping up

Screenshotting every page of a website is always two steps: enumerate the routes, then loop them through a capture. Consumer extensions cover the single page in front of you; a whole-site audit needs the route list (from a sitemap or a crawl you run) and a loop that turns each URL into a hosted image. Point the loop at your list, collect the image URLs, and you have a repeatable visual audit that runs anywhere, including CI and serverless. To automate it end to end, see Grabbit's automated screenshots.

---

Originally published on the Grabbit blog.