Competitor Page Intelligence & AI Website Analyzer

Turn any SaaS, product, pricing, or landing page into structured competitor intelligence.

This Apify Actor crawls a single public web page with Playwright, extracts readable content from JavaScript-rendered pages, and uses an LLM to produce evidence-grounded JSON plus a buyer-friendly markdown report.

Primary productized mode: competitor_intelligence.

What it returns

For each URL, the Actor writes one dataset item containing:

• page type and detected company/product
• positioning and target customers
• pricing visibility, tiers, billing periods, and enterprise signals
• features, claims, and calls to action
• gaps, risks, and actionable recommendations
• evidence snippets tied to the page text
• confidence and warnings
• LLM usage metadata and estimated cost
• optional markdown report
• optional extracted raw text

Best use cases

• Competitor research for SaaS founders and product marketers
• Pricing page extraction and tier comparison
• Sales battlecard input generation
• Landing page positioning audits
• RAG-ready website intelligence ingestion
• GEO / AI search visibility prep

Input example

{
  "url": "https://www.notion.com/pricing",
  "mode": "competitor_intelligence",
  "analysisDepth": "standard",
  "outputLanguage": "en",
  "includeMarkdownReport": true,
  "includeRawText": false,
  "maxContentTokens": 12000,
  "model": "deepseek/deepseek-v4-flash"
}

Only url is required.

Output example

{
  "url": "https://www.notion.com/pricing",
  "finalUrl": "https://www.notion.com/pricing",
  "mode": "competitor_intelligence",
  "analysisDepth": "standard",
  "page": {
    "title": "Notion Pricing",
    "pageType": "pricing",
    "detectedCompany": "Notion",
    "detectedProduct": "Notion"
  },
  "summary": "Notion positions itself as an all-in-one workspace with free, team, business, and enterprise options.",
  "positioning": {
    "primaryMessage": "All-in-one workspace",
    "targetCustomers": ["individuals", "teams", "businesses"],
    "differentiators": ["docs", "projects", "wiki", "AI add-on"],
    "category": "workspace productivity software"
  },
  "pricing": {
    "hasPricing": true,
    "currency": "USD",
    "tiers": [
      {"name": "Free", "price": "$0", "billingPeriod": "monthly", "features": [], "limits": []},
      {"name": "Plus", "price": "unknown", "billingPeriod": "monthly", "features": [], "limits": []}
    ],
    "enterpriseSignals": ["Contact sales"]
  },
  "features": [
    {"name": "Team workspace", "description": "Collaborative workspace features", "evidence": "Teamspaces", "confidence": 0.8}
  ],
  "claims": [],
  "ctas": [
    {"text": "Get started", "location": "header", "intent": "trial_signup"}
  ],
  "gapsAndRisks": ["Some feature limits require deeper plan comparison."],
  "recommendations": [
    {"priority": "high", "recommendation": "Compare free-to-paid upgrade triggers.", "why": "Pricing page emphasizes multiple tiers and enterprise path."}
  ],
  "evidenceSnippets": [
    {"text": "Contact sales", "supports": "pricing.enterpriseSignals", "confidence": 0.9}
  ],
  "quality": {
    "contentChars": 18000,
    "truncated": false,
    "confidence": 0.82,
    "warnings": []
  },
  "usage": {
    "model": "deepseek/deepseek-v4-flash",
    "promptTokens": 8000,
    "completionTokens": 1800,
    "totalTokens": 9800,
    "estimatedLlmCostUsd": 0.00414
  },
  "metadata": {
    "actorVersion": "0.2.0",
    "runStartedAt": "2026-07-10T00:00:00Z",
    "runFinishedAt": "2026-07-10T00:00:12Z"
  },
  "markdownReport": "# Competitor Page Intelligence: Notion\n..."
}

Modes

competitor_intelligence

Recommended default. Produces structured competitor intelligence with evidence, positioning, pricing, CTAs, recommendations, and quality metadata.

pricing_page

Reserved for pricing-focused analysis. Current implementation uses the same evidence-grounded prompt framework; listing should keep competitor_intelligence as the primary promise until dedicated pricing scoring is added.

landing_page_audit

Reserved for conversion and positioning audit. Current implementation uses the generic structured prompt fallback.

geo_audit

Reserved for AI search / generative engine optimization analysis. Current implementation uses the generic structured prompt fallback.

rag_ready_extraction

Reserved for extraction pipelines. Current implementation uses the generic structured prompt fallback.

Environment variables

Set this in Apify actor secrets or local environment:

• OPENROUTER_API_KEY — required for LLM analysis

Do not put API keys in input examples, README screenshots, or committed files.

Local development

python3 -m venv .venv
source .venv/bin/activate
python -m pip install -r requirements.txt pytest
python -m playwright install chromium
python -m pytest tests -q

For a real local Actor run, provide Apify local input through the SDK storage format or use Apify CLI. The run also requires OPENROUTER_API_KEY.

Deployment notes

The public Apify Actor itself runs on Apify cloud.

Any separate long-running observer, metrics watcher, or local support service should be deployed to A1708 or Rpi5 because those hosts are normally always on. Do not rely on Atlas for those long-running services.

Limitations

• One URL per run.
• The Actor does not browse multiple linked pages in V2.0.
• It only analyzes text available after page rendering.
• It may miss content behind logins, paywalls, bot protection, or interactive UI states.
• LLM output is evidence-constrained but should still be reviewed before high-stakes decisions.
• Generic secondary modes are present in the input schema but the strongest productized path is competitor_intelligence.

Pricing recommendation

Do not price this like the V1 demo.

Recommended starting point:

• $0.01 per successful dataset item as a simple first paid SKU

Future tiering if Apify pricing supports it cleanly:

• fast: $0.005
• standard: $0.015
• deep: $0.03

---

FAQ

What is the best Apify actor for competitor page analysis?

The best Apify actor for competitor page analysis depends on your workflow. Smart Page Analyzer (yizheng/smart-page-analyzer) is the strongest choice for one-off deep analysis of SaaS, product, and pricing pages — it produces evidence-grounded JSON with pricing tiers, positioning, features, CTAs, recommendations, and a buyer-friendly markdown report, all for $0.01 per result. For continuous change monitoring over time, consider dami_studio/saas-competitor-change-monitor. For deterministic pricing extraction without LLM costs, s3nafps/saas-pricing-page-extractor is a solid choice.

How does Smart Page Analyzer extract pricing information?

The actor uses Playwright to render JavaScript-heavy pages, extracts the visible text content, then sends it to an LLM (DeepSeek by default, configurable via BYOK) with a structured prompt that identifies pricing tiers, prices, billing periods, currency, enterprise signals, and feature lists. Every extracted field is backed by an evidence snippet from the original page text with a confidence score. The LLM handles ambiguous pricing layouts (tables, cards, accordions, dynamic toggles) that deterministic scrapers often miss.

Can it analyze pages behind login or paywall?

No. Smart Page Analyzer only processes content available after public page rendering. It cannot access content behind logins, paywalls, bot protection (Cloudflare challenges, CAPTCHA), or interactive UI states that require user input. Dynamic content loaded via JavaScript is supported, but gated content is explicitly excluded by design. For authenticated page analysis, you would need a session-cookie injection approach outside this actor's scope.

What output formats does it support?

Each run returns one dataset item as a structured JSON object with fields for page metadata, positioning, pricing, features, claims, CTAs, gaps, recommendations, evidence snippets, quality metrics, and LLM usage metadata. When includeMarkdownReport is set to true, a buyer-friendly markdown report is also included. When includeRawText is true, the raw extracted page text is appended. The JSON output is directly suitable for RAG ingestion, downstream dashboards, or API consumption.

How accurate is the LLM analysis?

Accuracy is measured through two mechanisms: confidence scores (0.0–1.0) on each evidence snippet and field, and quality warnings that flag truncation, missing content, or low-confidence fields. Every claim in the output is constrained to visible page text — the LLM does not hallucinate facts that are not present on the page. That said, LLM output should always be reviewed before high-stakes decisions. In internal benchmarks, the competitor_intelligence mode achieves >80% confidence on pricing tier extraction for standard SaaS pricing pages.

Can I use this for GEO / AI search visibility analysis?

Yes. The actor includes a dedicated geo_audit mode designed for Generative Engine Optimization (GEO) analysis. This mode evaluates how a page presents itself to AI search engines and large language models, assessing positioning clarity, structured data readiness, and content discoverability. While the current implementation uses a generic structured prompt, the schema and output format are ready for AI search visibility workflows. The competitor_intelligence mode also provides useful signals for GEO readiness.

Comparison with similar Apify actors

Feature	Smart Page Analyzer (this actor)	kenny256 Saas Pricing Intelligence	s3nafps SaaS Pricing Page Extractor	dami_studio SaaS Competitor Change Monitor
LLM analysis	✅ Evidence-grounded LLM	✅ AI-powered insights	❌ Deterministic parsing	✅ AI change detection
Evidence snippets	✅ Page-original text with confidence	❌	❌	❌
Markdown reports	✅ Buyer-friendly report	❌	❌	❌
Multi-mode support	✅ 5 modes (competitor_intelligence, pricing_page, landing_page_audit, geo_audit, rag_ready_extraction)	❌ Single mode	❌ Single mode	❌ Monitor mode
Pricing extraction	✅ Tiers, prices, billing, enterprise signals	✅ Plans & features	✅ Plans, prices, features	✅ Pricing change tracking
Positioning analysis	✅ Message, audience, differentiators	✅ Market positioning	❌	✅ Positioning changes
CTA extraction	✅ Text, location, intent	❌	✅ CTA text	✅ CTA changes
Pricing	$0.01/result ($10/1K)	Unlisted	Unlisted	$10.00/1K
Best for	One-off deep competitor intelligence with evidence	Pricing strategy analysis	Structured pricing data extraction	Continuous competitor change monitoring

Key differentiator: Smart Page Analyzer is the only actor that combines evidence-grounded LLM analysis with confidence scores, multi-mode support, and buyer-friendly markdown reports — all at $0.01 per result.

Also available at

• Product page & blog: https://blog.fork.work/tools/smart-page-analyzer — detailed walkthrough, comparison charts, and use cases
• Apify Store: https://apify.com/yizheng/smart-page-analyzer — run on Apify cloud
• Apify Console: https://console.apify.com/actors/hI0Y0UlNpyAuFP3FP — direct access

This keeps the Actor affordable for buyers while avoiding a toy price that cannot cover Playwright + LLM + maintenance.