Public Screenshot Visual Readiness Agent

Analyze public pages for screenshot and visual regression readiness using static HTML evidence. The Actor does not take screenshots or run a browser; it helps agents decide whether a page looks simple enough for a basic capture or needs waits, scrolling, consent handling, iframe/media handling, or a browser renderer.

Use it before spending money on browser-based screenshots, visual regression runs, or QA capture jobs. It turns a public page into compact capture guidance without executing JavaScript or touching private sessions.

How To Use It

For a first pass, provide public page URLs and leave previous records empty:

{
  "screenshotPageUrls": ["https://www.python.org/"],
  "focusTerms": [],
  "previousScreenshotReadinessRecords": [],
  "requestTimeoutSecs": 20
}

For repeat monitoring, pass previous dataset records from this Actor in previousScreenshotReadinessRecords. If the screenshotReadinessHash is unchanged for the same normalized page, the result is written without charging the useful event.

Optional focusTerms can narrow billing to evidence that mentions a specific capture concern, such as cookie, modal, lazy, iframe, video, React, viewport, animation, or hero.

What It Returns

• Screenshot readiness type: static capture ready, JavaScript render risk, overlay/cookie risk, embed/media heavy, auth/paywall risk, or needs capture waits.
• Viewport, responsive, visual asset, lazy-load, iframe, media, overlay, animation/stability, and JavaScript framework hints.
• Evidence snippets, sanitized evidence URLs, diagnostics, readiness/risk/ confidence/completeness scores, a recommended capture plan, and a stable screenshotReadinessHash.

Pricing Behavior

The Actor uses pay-per-event pricing and only charges the useful result event when it finds useful, new or changed screenshot-readiness evidence.

Paid event:

• useful-screenshot-readiness-result

It does not charge the useful event for:

• invalid or private-network URLs;
• failed fetches;
• query strings, fragments, credentials, login/account/admin paths, or token-like path segments;
• pages with too little public visual-readiness evidence;
• focusTerms misses;
• unchanged records when you pass a previous screenshotReadinessHash;
• duplicate normalized URLs in the same run.

Default dataset item charging is intentionally not configured.

Scope And Safety

• Public HTTP/HTTPS HTML or text pages only.
• JavaScript is not executed and screenshots are not taken.
• Forms are not submitted, banners are not clicked, and no private sessions are used.
• Image, iframe, media, stylesheet, and canonical URLs are recorded only as sanitized evidence links and are not fetched.
• Private-network hosts, credentials, query strings, fragments, path parameters, login/account paths, and token-like paths are rejected.

Best Use Cases

• Preflight screenshot or visual regression jobs before paying for browser rendering.
• Detect likely cookie/modal, lazy-load, iframe, media, or JS framework complications.
• Monitor when a page's visual-capture readiness changes.
• Feed AI QA pipelines with compact source-linked capture guidance.

Not For

This Actor is not a screenshot API, browser renderer, pixel-diff engine, uptime monitor, or accessibility audit. It reads public page HTML and returns readiness evidence so another workflow can decide how to capture the page.