Pricing & Feature Change Monitor

Pricing page change monitor (Cheerio) for public SaaS pages — watch competitor pricing, run feature page change detection, or use as a SaaS pricing page monitor / feature changelog tracker. Fetches HTML without a browser, normalizes text, compares SHA-256 to the last snapshot, and writes baseline | unchanged | changed | error dataset rows.

Store: apify.com/frqnk/pricing-feature-change-monitor
Actor: frqnk/pricing-feature-change-monitor · ID: wlL2c2qokY0BzCe8W

For AI agents

This section is for Cursor agents, Claude (Computer Use / Desktop MCP), OpenAI agents, Apify MCP clients, and other runners that discover Actors via search-actors or a pinned MCP tool list.

Actor ID: frqnk/pricing-feature-change-monitor
Markdown: https://apify.com/frqnk/pricing-feature-change-monitor.md
Platform quickstart: https://apify.com/agents.md

What this Actor does for agents

Fetches public pricing/feature pages with Cheerio (no browser), normalizes text, compares SHA-256 to the last snapshot in a named KV store, and emits dataset rows: baseline | unchanged | changed | error. Agents should schedule re-runs with the same stateStoreName and branch only when status === "changed".

When to use / when not

Use	Do not use
Public pricing / feature marketing pages	Login, cookies, authenticated dashboards
Structured change signal for schedules	Heavy SPA shells that need Playwright
Text diffs of plan/feature copy	Visual/screenshot QA
Lightweight PPE monitoring	Full product-catalog scrapes

Cost disclosure (required before first paid run)

Pay-per-event. State this to the user once per session before the first paid run; wait for go-ahead. Cap with callOptions.maxTotalChargeUsd (MCP) — never put caps in input.

Event	Price	When
url-check	$0.002	Successful check (unchanged or changed)
change-detected	$0.04	Meaningful hash change
Baseline / error	$0	First-seen or failure

Estimate: (N_checks × $0.002) + (N_changes × $0.04).

MCP setup

Default discovery (search Store, then call):

{
  "mcpServers": {
    "apify": {
      "type": "http",
      "url": "https://mcp.apify.com"
    }
  }
}

Pin this Actor (recommended for pricing-watch skills):

{
  "mcpServers": {
    "apify": {
      "type": "http",
      "url": "https://mcp.apify.com/?tools=actors,docs,frqnk/pricing-feature-change-monitor"
    }
  }
}

OAuth on first connect, or Authorization: Bearer <APIFY_TOKEN>. CLI: apify mcp install cursor --tools actors,docs,frqnk/pricing-feature-change-monitor.

Anonymous discovery only (no runs):
https://mcp.apify.com/?tools=search-actors,fetch-actor-details,search-apify-docs,fetch-apify-docs

Call example (MCP tool sequence)

1. fetch-actor-details → frqnk/pricing-feature-change-monitor (or search-actors with keywords pricing feature change monitor)
2. Tell user: PPE $0.002 / url-check, $0.04 / change-detected; estimate for N URLs; wait for go-ahead
3. call-actor:

{
  "actor": "frqnk/pricing-feature-change-monitor",
  "input": {
    "urls": ["https://stripe.com/pricing", "https://github.com/pricing"],
    "profile": "pricing",
    "ignoreNoise": true,
    "emitBaseline": true,
    "stateStoreName": "pricing-feature-change-state"
  },
  "callOptions": { "maxTotalChargeUsd": 1 }
}

4. get-dataset-items with the returned datasetId
5. Filter status === "changed"

call-actor returns run metadata and storage IDs, not dataset rows. Always fetch items.

Success criteria

Check	Pass condition
Run	Status SUCCEEDED
Data	Dataset items loaded
Monitor loop	Same stateStoreName across runs; act on changed only
Cost	User approved estimate before first paid run; maxTotalChargeUsd set
Failure	On FAILED / TIMED-OUT / ABORTED, read statusMessage / logs; do not retry blindly

Permissions

Limited: public HTTP(S) fetches + run dataset + named KV store. No login automation.

Skill paste

See companion file MCP_SKILL_SNIPPET.md in the agent-targeted marketing pack, or paste the one-liner:

For public pricing/feature page change watches: Apify Actor frqnk/pricing-feature-change-monitor — PPE $0.002/url-check, $0.04/change-detected; baselines+errors free. State cost before first paid run. Cheerio only.

Use cases — SaaS pricing page monitor & competitor pricing watch

• Watch competitor pricing — notice when a SaaS rival updates plan names, prices, or packaging copy
• Feature changelog tracker / feature page change detection — catch marketing-site feature edits without an official feed
• Pricing page change detection — scheduled hash checks; emit diffs only when content meaningfully changes
• Agent / MCP automation — Schedules + filter changed rows into Slack, email, or another tool

Limitations (honest)

• Public HTML only — no logins, cookies, paywalls, or authenticated dashboards
• Cheerio change monitor (static HTML) — best on server-rendered pricing and feature pages. Heavy SPA shells need a browser Actor; this one does not use Playwright
• Noise filters (ignoreNoise) reduce false positives from timestamps and counters but are not perfect
• Diffs are text heuristics (changeType / severity), not visual QA or pixel comparison

How pricing page change detection works

1. Fetch each URL with Crawlee CheerioCrawler
2. Normalize HTML (script/style stripped; optional contentSelector; optional noise stripping)
3. Hash normalized text (SHA-256) and load the previous snapshot from a named Key-Value store
4. Emit a dataset row and charge PPE only when the rules below say so

Situation	Dataset	PPE
First seen, emitBaseline: true	baseline row	No charge
First seen, emitBaseline: false	No row (state saved)	No charge
Same hash	unchanged	url-check
Different hash	changed + diff snippets	url-check + change-detected
Fetch / process failure	error	No charge

Demos & Store example tasks: keep emitBaseline: true so the first run produces rows quickly.
Production schedules on the same URLs: you can set emitBaseline: false if you only want unchanged / changed after the first silent snapshot.

Quick start input

{
  "urls": [
    "https://stripe.com/pricing",
    "https://github.com/pricing"
  ],
  "contentSelector": "main",
  "profile": "pricing",
  "ignoreNoise": true,
  "minChangePercent": 1,
  "stateStoreName": "pricing-feature-change-state",
  "emitBaseline": true,
  "maxConcurrency": 5
}

Field	Default	Notes
urls	(required)	Public pricing / feature page URLs to monitor
contentSelector	main	Region hashed; falls back to body
profile	pricing	Heuristic for changeType / severity (pricing, feature, generic)
ignoreNoise	true	Strip dates / “Updated: …” / counters
minChangePercent	1	Influences severity banding
stateStoreName	pricing-feature-change-state	Named KV for snapshots across runs — reuse across schedules
emitBaseline	true	Emit first-seen rows (never charged)
maxConcurrency	5	Parallel Cheerio fetches

Open the Actor in Console, paste URLs, click Start, or duplicate a published Example task (Stripe / GitHub / multi-URL SaaS pack).

Output example

{
  "url": "https://stripe.com/pricing",
  "status": "changed",
  "changed": true,
  "checkedAt": "2026-09-20T15:00:00.000Z",
  "previousCheckedAt": "2026-09-19T15:00:00.000Z",
  "contentHash": "452029c4…",
  "previousHash": "a1b2c3d4…",
  "changeType": "pricing",
  "severity": "major",
  "addedText": ["Pro plan costs $49 per month."],
  "removedText": ["Pro plan costs $29 per month."],
  "diffSnippet": "- Pro plan costs $29 per month.\n+ Pro plan costs $49 per month.",
  "httpStatus": 200,
  "errorCode": null
}

status: baseline | unchanged | changed | error

Pricing (pay-per-event / PPE)

Event	Price	When
url-check (primary)	$0.002	Successful fetch + compare (unchanged or changed)
change-detected	$0.04	Meaningful hash change
Baseline / error	$0	First-seen or failure

Platform apify-actor-start may apply per Apify PPE rules. Agents: disclose costs and wait for go-ahead before the first paid run; cap with callOptions.maxTotalChargeUsd.

FAQ

Can an AI agent use this as a pricing page monitor without a browser?

Yes for public, server-rendered HTML. It is a Cheerio change monitor — no Playwright — so SPA-heavy pages with empty initial HTML may return thin or empty content.

Can I watch competitor pricing behind a login?

No. Public pages only. Login walls, cookies, and authenticated dashboards are out of scope.

How is feature page change detection different from a full changelog product?

It tracks public page text via hash compare and short diffs. It does not scrape private release notes, admin panels, or visual layouts.

What does PPE charge on unchanged checks?

Successful compares charge url-check ($0.002). Only meaningful hash changes also charge change-detected ($0.04). Baselines and errors are free of those events.

Permissions

Limited permissions: public HTTP(S) fetches plus the run dataset and named KV store. No login automation, no residential-proxy requirement for typical public marketing pages.

Support

Issues and feature requests: open a discussion on the Store page or contact the publisher via Apify.

License

Apache-2.0