Evidence-Grade Web Page Diff

Evidence-Grade Web Page Diff captures a stable, normalized text snapshot of one public HTTPS page. A later run can compare that page with a caller-supplied snapshot and return bounded paragraph additions and removals for policy pages, terms, disclosures, public notices, or other text where a deterministic record matters.

This is a stateless page-diff product, not a persistent monitoring service. API-IFY does not retain named baselines or schedule future comparisons; callers retain and resubmit the snapshot, optionally using Apify schedules.

Input

Provide a public HTTPS url. To compare a page, also provide the complete WEBSITE_SNAPSHOT object from an earlier run as baseline. The Actor verifies that baseline text matches its SHA-256 hash and that the prior source matches the final fetched page. Set includeNormalizedText when the dataset item should include the normalized body in addition to the reusable snapshot.

{
  "url": "https://example.com/policy",
  "includeNormalizedText": false
}

Only HTTPS port 443 is supported. URL credentials, private targets, mixed public/private DNS responses, and redirects to private destinations are rejected. Before fetching the page, the Actor evaluates robots.txt for its configured user agent. A disallow rule produces a restricted, uncharged result.

Output

Every charged successful run writes WEBSITE_SNAPSHOT to the default key-value store before publishing the dataset result. It contains schemaVersion, final sourceUrl, capture time, content hash, and normalized text. Normalized text is limited to 2,000,000 characters so every emitted snapshot remains valid as a future baseline; larger pages are rejected without charge. The default dataset receives a stable API-IFY envelope with the current and baseline hashes, change state, magnitude, classification, bounded additions and removals, and analyzer provenance. Navigation, scripts, styles, templates, footers, common consent banners, and hidden content are removed before comparison.

A new snapshot produces a result projection like this:

{
  "status": "success",
  "result": {
    "classification": "snapshot",
    "currentHash": "af0270fac35233692c0ada7bb34efaf6bf81304c0d738eb645366f23d4fff70e",
    "snapshotKey": "WEBSITE_SNAPSHOT"
  }
}

See examples/sample-input.json and examples/sample-output.json for complete schema-valid contracts.

RUN_SUMMARY reports result status, HTTP requests and retries, bytes processed, duration, budget skips, and charged events. Invalid baselines, unsupported pages, robots restrictions, unsafe targets, and fetch failures produce explicit uncharged dataset records.

Pricing

A new snapshot charges one page-snapshot event. A successful baseline comparison charges one page-diff event. The Actor checks the remaining event budget before network retrieval. Failed or restricted results are not charged. Current event prices are displayed by Apify.

Privacy and operational behavior

API-IFY keeps no separate website database. Data is written only to the run's default Apify dataset and key-value store, where the caller controls retention and access. Published structures are sanitized for authorization values, API keys, tokens, passwords, secrets, private keys, and sensitive URL query parameters. The full request URL exists only in memory while retrieval is performed.

To run recurring comparisons, configure an Apify schedule and pass the previously retained snapshot into the next run. For support, provide the run ID, result itemId, and public page URL without credentials. Contact Sentinel Signal Systems through the support link on the Actor page.