Contact Page & Sales NG Policy Detector

An ultra-fast, lightweight Apify Actor designed to crawl business websites, locate Contact Pages, detect Sales NG policies (bans on commercial outreach/pitches), evaluate Contact Forms, and identify CAPTCHA protection.

Engineered specifically for B2B cold outreach compliance (with deep multi-level Japanese linguistic analysis) and bulk processing of up to 1,000 websites per batch.

---

🚀 Key Features

• 🎯 Smart Contact Page Finder: Uses a multi-criteria scoring algorithm (URL patterns, anchor text, header/footer context, onclick handlers) to accurately pinpoint contact pages.
• 🚫 Sales NG Policy Detection: Multi-level semantic analysis to identify explicit bans on sales, pitches, or unsolicited emails (e.g., 営業目的...ご遠慮ください).
• 📝 Form & Field Analytics: Detects active HTML <form> elements and counts interactive user fields (input, textarea, select).
• 🛡️ CAPTCHA Recognition: Identifies Google reCAPTCHA, hCaptcha, Cloudflare Turnstile, and Japanese verification systems (画像認証).
• ⚡ Resource & Speed Optimized: Blocks images, media, fonts, and tracking scripts to run with ultra-low RAM/CPU overhead and maximum execution speed.

---

📥 Input Configuration

The Actor expects a JSON input containing a direct link to a CSV file hosted on an Apify Key-Value Store (or any public HTTP URL).

Input JSON Example

{
  "inputFile": "https://api.apify.com/v2/key-value-stores/YOUR_STORE_ID/records/input.csv",
  "maxPages": 20
}

Input Parameters

Parameter	Type	Required	Default	Description
inputFile	String	Yes	-	Public URL to the CSV file stored in an Apify Key-Value Store.
maxPages	Integer	No	20	Maximum pages processed per crawl execution.

CSV File Requirements

The CSV file must contain a header with exactly two columns: id (integer) and url (valid website URL). Supports up to 1,000 URLs per run.

id,url
1,https://example.com
2,https://www.beautygarage.co.jp

---

📤 Output Data Format

The results are stored in the default Apify Dataset as JSON items:

Sample Dataset Output

{
  "id": 2,
  "url": "https://www.beautygarage.co.jp",
  "contactPage": "https://www.beautygarage.co.jp/inquiry/",
  "contactPageFound": true,
  "salesStatus": "RESTRICTED",
  "restrictionScore": 100,
  "restrictionReason": "営業目的 + ご遠慮ください | 営業目的のお問い合わせはご遠慮ください。",
  "matchedSubjects": "営業目的",
  "matchedPredicates": "ご遠慮ください",
  "hasForm": true,
  "formCount": 1,
  "fieldCount": 5,
  "captcha": true,
  "captchaType": "reCAPTCHA",
  "captchaCount": 1,
  "error": null
}

Dataset Field Descriptions

Field	Type	Description
id	Integer	The unique ID supplied in your input CSV.
url	String	The target homepage URL.
contactPage	String / Null	Discovered URL of the contact page.
contactPageFound	Boolean	true if a contact page was successfully identified.
salesStatus	String	Sales policy status: ALLOW, WARNING, RESTRICTED, or ERROR.
restrictionScore	Integer	Risk score (0 = Safe, 30–40 = Warning, 100 = Explicitly Restricted).
restrictionReason	String	Contextual proof (Subject + Predicate + Matched sentence).
matchedSubjects	String	Matched commercial outreach subjects (e.g., 営業目的, セールス).
matchedPredicates	String	Matched rejection/restriction terms (e.g., ご遠慮ください, お断り).
hasForm	Boolean	true if interactive contact forms exist on the page.
formCount	Integer	Total count of <form> tags found.
fieldCount	Integer	Count of usable input fields (input, textarea, select).
captcha	Boolean	true if CAPTCHA mechanisms were detected.
captchaType	String	Type of CAPTCHA detected (reCAPTCHA, hCaptcha, Turnstile, CAPTCHA).
captchaCount	Integer	Occurrences of CAPTCHA indicators in page sources.
error	String / Null	Error message if the navigation or crawl failed.

---

🚦 Understanding salesStatus

• RESTRICTED (Score: 100): The target page explicitly forbids commercial pitches or sales forms. Do not send outreach emails or submit forms here to avoid domain blacklisting.
• WARNING (Score: 30–40): Isolated sales or rejection keywords were detected, but no definitive restriction sentence was formed. Proceed with caution.
• ALLOW (Score: 0): No restriction policies detected. Safe for outreach.
• ERROR: Website failed to load or timed out during navigation.

---