Website Change Monitor: Detect Page Changes, Diff and Webhook avatar

Website Change Monitor: Detect Page Changes, Diff and Webhook

Pricing

from $15.00 / 1,000 page checkeds

Go to Apify Store
Website Change Monitor: Detect Page Changes, Diff and Webhook

Website Change Monitor: Detect Page Changes, Diff and Webhook

Website change detection for pricing pages, terms, changelogs and competitor sites. Filters noise (dates, counters, rotating banners), watches a CSS selector, shows a clean diff and calls your webhook on change. Respects robots.txt. Failed checks are free.

Pricing

from $15.00 / 1,000 page checkeds

Rating

0.0

(0)

Developer

Tan Solutions

Tan Solutions

Maintained by Community

Actor stats

0

Bookmarked

2

Total users

1

Monthly active users

3 days ago

Last modified

Categories

Features

Limited permissions

Share

Website change detection for pricing pages, terms, changelogs and competitor sites. Give it a list of URLs and a schedule, and every run tells you which pages changed since the last run, with a clean diff of what was added and removed. Noise (dates, times, counters, cache-busting tokens, rotating banners) is filtered out, you can watch just one part of a page with a CSS selector, and your webhook is called only when something really changed. Failed checks are never charged.

Quick API and schedule example

Run it from code:

curl -X POST "https://api.apify.com/v2/acts/tan-solutions~website-change-monitor/run-sync-get-dataset-items?token=$APIFY_TOKEN" \
-H "Content-Type: application/json" \
-d '{"startUrls": [{"url": "https://example.com/pricing", "selector": "main"}], "stateKey": "competitor-pricing", "webhookUrl": "https://hooks.example.com/changes"}'

Then put the same input on a schedule: Apify Console > Schedules > Create, add this Actor with the same input, and pick a time, for example 0 8 * * * (every day at 08:00). Every scheduled run compares with the run before it. Keep stateKey the same on every run of one watch list; a new stateKey starts a fresh list with its own baselines.

What it does

  • Fetches each URL over plain HTTP (no browser), extracts the main text of the page (or only the part your CSS selector matches) and compares it with the text saved at the last check.
  • One row per URL and run, with a status: changed, unchanged, new (first check, baseline saved) or failed (with an error code and a plain message).
  • For a change: a unified diff (capped at maxDiffChars), the added and removed text as short snippets, a changeScore from 0 to 1 (the share of the text that changed) and when the page was checked before.
  • Noise filter (on by default): dates, clock times, "3 hours ago", view, like and follower counters and cache-busting tokens are ignored when comparing, and carousels, sliders, tickers, ads, cookie and consent banners, pop-ups and chat widgets are removed before the text is read. Prices, plan limits, version numbers and percentages are never treated as noise: a pricing page that goes from $29 to $39 reads changed. Add your own ignorePatterns (regular expressions) for anything else that keeps changing.
  • Webhook: with webhookUrl set, the run sends one POST with all changed pages at the end, and only when at least one page changed.
  • Baselines live in a named key-value store in your own account (tansolutions-changes-<stateKey>), so they survive between runs and schedules. A failed check keeps the old baseline, so the next good check still reports the change.

Who it is for

  • Marketers and founders who watch competitors' pricing pages, plans, feature lists and landing pages.
  • Compliance, legal and procurement teams who track terms of service, privacy policies and supplier pages.
  • Developers who follow changelogs, docs and release pages, and AI agents that need "what changed since last time".

Quick start

  1. Add pages in Pages to watch, one per line. To watch only part of a page, set CSS selector (for all pages), or give one page its own selector in JSON: {"url": "https://example.com/pricing", "selector": "#plans"}.
  2. Give the list a Monitor name (stateKey), for example competitor-pricing.
  3. Click Start. The first run saves the baselines (new); every later run reports changed or unchanged.
  4. Schedule it (Console > Schedules) and add a Webhook URL if you want a POST on every change.

Input

Field (Console name)What it doesDefault
startUrls (Pages to watch)URLs, or {"url", "selector"} objects for a selector per pageone Apify docs page (demo)
mode (Mode)monitor compares with the last check; snapshot only saves new baselinesmonitor
selector (CSS selector (all pages))Watch only the part of each page this selector matcheswhole page
ignorePatterns (Ignore patterns)Up to 20 regular expressions; matching text is ignored when comparingnone
noiseFilter (Filter noise)Ignore dates, times, counters, cache-busting tokens and rotating bannerson
webhookUrl (Webhook URL)POST with the changed pages, sent only when something changednone
stateKey (Monitor name)Names the saved baselines (letters, digits, hyphens, at most 40)the task ID, else default
includeText (Include page text)Add the full compared text to every rowoff
includeLinks (Compare link targets)Also report a change when only a link target changedoff
maxDiffChars (Max diff length)Diff length per row, 500 to 50,000 characters5,000
minRecheckSecs (Minimum re-check interval)A page checked more recently than this is skipped, free; 60 s to 7 days300
maxUrls (Max URLs)Pages per run, 1 to 1,000100
timeoutSecs (Timeout per page)5 to 120 seconds30
maxConcurrency (Parallel pages)Pages fetched at once, 1 to 205
proxyConfiguration (Proxy)Optional Apify Proxyoff
{
"startUrls": [
{ "url": "https://example.com/pricing", "selector": "#plans" },
{ "url": "https://example.com/terms" }
],
"stateKey": "competitor-pricing",
"ignorePatterns": ["Order #\\d+"],
"webhookUrl": "https://hooks.example.com/changes"
}

Opening hours and timetables: the noise filter ignores clock times ("9:00") and weekday names in dates. To watch a page where those are the content (opening hours, timetables, event times), set noiseFilter: false and use ignorePatterns for the parts that really are noise.

Output example

One dataset item per URL and run. Views: Overview and Diffs. A changed pricing page:

{
"url": "https://example.com/pricing", "finalUrl": "https://example.com/pricing", "selector": "#plans",
"status": "changed", "title": "Pricing", "changeScore": 0.0412,
"diff": "@@ -3,3 +3,3 @@\n ## Pro\n-$29 per month\n+$35 per month\n Unlimited projects",
"diffTruncated": false, "addedText": ["$35 per month"], "removedText": ["$29 per month"],
"addedLines": 1, "removedLines": 1, "lineCount": 24, "contentHash": "5c1f0e9a2b7d4c31",
"text": null, "httpStatus": 200,
"checkedAt": "2026-11-14T08:00:12.511Z", "previousCheckedAt": "2026-11-13T08:00:10.204Z",
"note": null, "error": null
}

A failed row has status: "failed" and an error, for example {"code": "SELECTOR_NOT_FOUND", "message": "the CSS selector \"#plans\" matched nothing on the page (the page layout may have changed)"}. Error codes: ROBOTS_DISALLOWED, BLOCKED (CAPTCHA or bot wall), HTTP_ERROR (4xx/5xx), TIMEOUT, DNS_ERROR, CONNECTION_FAILED, TOO_MANY_REDIRECTS, INVALID_URL, URL_NOT_ALLOWED, UNSUPPORTED_CONTENT_TYPE, CONTENT_TOO_LARGE, EMPTY_CONTENT, SELECTOR_NOT_FOUND, RATE_LIMITED (checked again too soon, or the site keeps answering 429), BLOCKED_BACKOFF (the page blocked an earlier check: see below), MAX_PAGES_REACHED (over maxUrls), INVALID_INPUT (a bad URL, or the same URL and selector twice in one input: it is checked once), INTERNAL_ERROR. If a page's HTML is too broken to convert, the row still succeeds with the page's plain text and parseFallback: true.

The webhook body (one POST per run, only if a page changed; at most 100 changes, each diff cut at 2,000 characters; the full diffs are in the dataset):

{
"event": "website.changed", "stateKey": "competitor-pricing", "runId": "a1B2c3D4e5F6g7H8i",
"checkedAt": "2026-11-14T08:00:15.002Z", "changedCount": 1, "truncated": false,
"changes": [{ "url": "https://example.com/pricing", "finalUrl": "https://example.com/pricing", "selector": "#plans",
"title": "Pricing", "changeScore": 0.0412, "diff": "...", "addedText": ["$35 per month"],
"removedText": ["$29 per month"], "checkedAt": "2026-11-14T08:00:12.511Z",
"previousCheckedAt": "2026-11-13T08:00:10.204Z" }]
}

Pricing

You pay per page check (the page-check event): every changed, unchanged or new row, including the first check that saves a baseline and every check in snapshot mode. A change costs nothing extra, and so does the webhook.

Apify planPrice per 1,000 page checks
Free$40
Bronze$28
Silver$20
Gold$15

TODO-CEO: these are proposed prices, not yet confirmed.

Free: every failed row (robots.txt, bot walls, 4xx/5xx, timeouts, selector misses, pages checked again too soon, duplicates, URLs over maxUrls). A failed or unpaid check keeps the old baseline. If your spending limit is reached, the run checks only the pages it can still pay for, then stops cleanly and says so; no saved baseline is lost. Apify's standard Actor start event is charged when a run starts, at the platform's default price.

Limits

  • Plain HTTP only, no browser: pages that build their content with JavaScript show little or no text. Pick server-rendered pages, or a selector that is in the page's HTML.
  • At most 1,000 URLs per run (default 100), and one check per URL every 60 seconds at the fastest (default 300 s).
  • Politeness: at most 2 requests at once and at least 1 second between requests to one site; Crawl-delay in robots.txt is honoured (a site asking for more than 30 s is skipped as RATE_LIMITED).
  • It obeys robots.txt (the TanForgeBot token, then *), never logs in and stops at CAPTCHAs and bot walls (BLOCKED). It uses no proxy unless you turn one on, and never rotates addresses to get past a block.
  • Pages behind bot walls or robots refusals fail free; a browser fallback is not in v1.
  • It identifies itself as TanForgeBot/1.0 (+https://apify.com/tan-solutions). Site owners can allow or block it in robots.txt with the TanForgeBot token.
  • Alerts go to your webhook (or Apify's own run notifications and integrations); there is no built-in email.
  • Pages over 5 million characters of HTML, or over 2 million characters of markup outside scripts and styles, fail free with CONTENT_TOO_LARGE.

FAQ

Why did a page read unchanged when its date changed? That is the noise filter at work: dates, times and counters are ignored. Turn noiseFilter off to compare the raw text.

My selector no longer matches. What happens? The row fails free with SELECTOR_NOT_FOUND and the old baseline is kept; fix the selector and the next run compares as before.

Can I check a page every minute? The fastest is once every 60 seconds per page (minRecheckSecs); a re-check inside that window is a free failed row with RATE_LIMITED.

A site blocked the check. What then? A bot wall or CAPTCHA, HTTP 401, 403 or 429, or a robots.txt disallow is a free failed row, and that page is not requested again for 24 hours (or minRecheckSecs, if longer). Runs in between list it as a free failed row with BLOCKED_BACKOFF and send no request; the baseline is kept. Network errors and 5xx answers are simply tried again on the next run.

I changed the selector or the ignore patterns. Do I get a fake change? No: that page starts a new baseline (new).

Can I call it from an AI agent, n8n, Make or Zapier? Yes, through the Apify API, SDKs, MCP or those integrations: pass startUrls and stateKey, read the dataset items, or receive the webhook.

Your responsibilities

You are responsible for the URLs you submit and for lawful use of the text you get back (copyright, the site's terms, any restrictions the site states, and privacy law for any personal data on the page). The Actor obeys robots.txt, does not log in and stops at CAPTCHAs and bot walls. It is provided without warranty; liability is capped at the fees you paid in the last 12 months. Questions and issues are answered within 14 days through the Actor's Issues tab.

Not affiliated with or endorsed by any website you monitor. All names belong to their owners.

Made by Tan Solutions.