Website Change Tracker

Know the moment any web page you care about changes — a regulator's new rule, a target account's press page, a vendor's terms, a competitor's pricing — without checking a single page yourself.

Pages change quietly: a regulator posts a new rule, a target account adds three senior roles, a vendor edits its terms, a competitor bumps a price — and you hear about it weeks later. Checking by hand doesn't scale.

So this Actor watches for you. Paste the pages to track; on every run it opens each one in a real browser, compares it to the last snapshot, and uses AI to tell you what changed and how big a deal it is (a 1–5 score). Every run that finds a change sends you a plain-English digest — in Slack, by email, or both — and adds to a growing Notion archive of every change. Slack, email, and Notion are optional.

One engine, many jobs. Pick a monitoring mode in What are you monitoring for? — general, competitor, sales signals, news, regulatory, vendor, hiring, brand, grants, or API changes — and the AI reads every change through that lens (a regulatory watch flags new rules and deadlines; a sales watch flags buying signals and outreach angles). Want the wording tuned even tighter? Add a custom write-up prompt and it reshapes how every summary reads, on top of the mode.

🧪 Early access. Priced lean — a small start fee (~$0.10/run), plus your Apify usage for the crawl, plus a small per-summary AI charge (free with your own AI key). No per-page or per-change markup. See Costs. Something off? Open an issue on the Issues tab. 🙏

Who it's for

• Competitive intelligence — a weekly "what moved across our competitor set" (pricing, packaging, product, partnerships) in one Slack channel.
• Sales & RevOps — an instant ping on a target account's funding, new partnership, marquee customer, or leadership change — with a ready outreach angle.
• Compliance & legal — the day a regulator posts a new rule, consultation, or deadline, or a standards body changes a requirement.
• Procurement & vendor risk — a heads-up when a critical vendor edits its terms, pricing, SLAs, security posture, or sub-processors.
• Recruiters & talent teams — careers pages watched for hiring sprees, senior roles, and new functions or regions.
• Content, SEO & PR — rivals' new posts, changelogs, and announcements — plus your own brand mentions on third-party pages — caught early.
• News & research — genuinely new items on the pages and topics you follow, minus the cosmetic edits.
• Grants & bid teams — new funding opportunities, tenders, and RFPs on funder and procurement portals, with deadlines and eligibility.
• Developers & platform teams — API references, docs, and changelogs watched for new, changed, deprecated, or breaking endpoints, parameters, and versions.
• Anyone else — leave it on General and it flags any meaningful change on any page you give it.

What you get

• Changes, read for you — each page is diffed against last time and summarized in a sentence ("Pro plan rose $49→$59 and an Enterprise tier was added"; "new clause added on data retention"), with a 1–5 significance score so cosmetic edits don't drown out real moves.
• A digest, not a list — each run that finds changes sends a grouped "what moved" rundown to Slack and/or your inbox, scannable in 30 seconds.
• Wording on your terms — pick a mode to set the analyst lens, then optionally add a custom write-up prompt to shape how every summary reads ("turn each signal into a one-line cold-email opener").
• You set the noise level — a significance threshold means you only hear about changes that matter (skip cosmetic edits); flip on the all-clear to even get a "nothing changed" ping on quiet runs.
• A Notion archive — every change becomes a row in a Website Changes database the Actor builds for you. Searchable, filterable, yours.
• An initial website analysis — the first time it sees a site, it writes a one-time profile (what the site does, positioning, pricing, product focus, hiring signals). It's saved to your dataset and reused as AI context on every later run, so each change is read in light of what that site actually does.

Set it up in 2 minutes

1. Add the pages, pick a mode. Paste the URLs, one per line — the only required field. A bare domain (example.com) works; a root URL watches the whole site (up to 50 pages/run); a path like example.com/pricing watches just that section. Then pick What you're monitoring for so the AI frames changes for your job — or leave it on General.
2. Pick your AI (optional). Default Claude Haiku 4.5 needs no setup; switch to Sonnet for sharper reads or GPT/Gemini with your own key — and optionally add a custom write-up prompt to shape how summaries read.
3. Choose where results go (optional). In Get your results, add your email, a Slack connector + channel, and/or a Notion connector for the archive — pick any or none (the structured dataset is always produced). Slack and Notion use MCP connectors you add once in Apify Console → Settings → Integrations → MCP Connectors (each field carries the one-line how-to).
4. Schedule it. On the Schedule tab, pick how often to check — hourly, daily, weekly, whatever matches how fast your pages move. The run is the schedule: every run that finds a change sends the digest.

Costs

Pay only when you run — no subscription, no per-result markup:

• Start fee — $0.05 per GB of memory, per run (so ~$0.10/run at the default 2 GB), billed even with your own AI key.
• Apify platform usage (compute + proxy) for the crawl — the main variable cost, since it renders each page in a real browser. Scales with pages tracked and run frequency.
• A small AI charge per summary — $0.02 (Haiku) or $0.06 (Sonnet), one flat charge per write-up (each change, the digest, each initial website analysis). $0 with your own Anthropic key, or with GPT/Gemini (you pay that provider directly).

The first run writes no change summaries (just a baseline per page, plus the initial website analysis if enabled), so it's mostly the start fee + crawl. Tips: track specific paths over whole sites, keep the default Haiku, and run daily, not hourly.

Use it from your AI assistant (MCP)

Run this Actor from any MCP client (Claude, Cursor, …) through the Apify MCP server (https://mcp.apify.com) with your Apify token — just ask "what's changed on the pages I'm watching?" and your assistant triggers a check and reads the changes straight back.

It has the history, too. The Actor keeps a baseline of every page between runs and writes each run's changes to its dataset, so your assistant can pull past runs — "what changed this week vs last?" — not just the latest check. Each run reports what's new since the previous snapshot, and older runs stay in the dataset.

FAQ

Why no changes on the first run? It stores a baseline first (you'll see a baseline row per page). Change detection starts on the second run, once there's something to compare against.

Why didn't my digest arrive? A digest is sent only on a run that actually finds a change at or above your notify threshold. A run that finds nothing (or only cosmetic edits below your threshold) stays silent — turn on "Tell me even when nothing changed" for an all-clear either way, and double-check your Slack channel / email address.

Do I have to set up the Notion database? No — the Actor builds it. By default it lands at your workspace root; paste a Notion page ID to keep it inside that page instead.

Can I use my own AI account? Yes. GPT/Gemini require your key; for Claude it's optional (blank = built-in key, or paste your Anthropic key to bill your own account and skip our AI charge).

A site is behind anti-bot protection? Switch Proxy to Residential. If a page still returns no content, the log flags a possible block.

Is this allowed? It reads publicly accessible pages, like a browser does. Review each target's Terms of Service, robots policy, and your local laws first, and don't collect personal or restricted data.

Bugs, requests, or custom tweaks? Open an issue on the Issues tab — we read every one. 🙏