✨ AI Ad Creative Intelligence API

Turn competitor ads into a clear creative strategy — not another spreadsheet of raw copy.

Analyze LinkedIn, TikTok, Meta, Google, or pasted ad records with GPT. For every ad, the Actor extracts the hook, offer, CTA, funnel stage, audience hypothesis, message angles, creative patterns, strengths, risks, evidence, and concrete experiments your team can run next.

Use it for competitor research, message mining, campaign audits, and creative-testing backlogs. Paste a few ads for an instant audit or connect an existing Apify dataset for batch analysis.

🎯 What you get

Every source ad becomes one structured, audit-ready result:

Output	Why it matters
primaryHook + hookType	See how the ad earns attention.
offer + cta	Compare conversion mechanics across competitors.
funnelStage	Separate awareness creative from conversion creative.
targetAudience	Capture evidence-based audience hypotheses.
angles + creativePatterns	Find messages worth adapting across a portfolio.
strengths + weaknesses	Understand what is clear, credible, or missing.
testIdeas	Turn research directly into an experimentation backlog.
complianceRisks	Flag quantitative, social-proof, or absolute claims for review.
evidence + confidence	Trace conclusions back to the supplied copy.
longevityDays	Use first/last-seen dates as an observable longevity signal.

The run summary also ranks recurring hook types, angles, and CTAs across the whole batch.

The Actor never invents ROAS, conversions, spend, or verified targeting. Audience and strategy fields are explicitly hypotheses based on the supplied creative.

🚀 Quick start

The default input is ready to run. Click Start for an immediate example, or paste your own ads:

{
  "ads": [
    {
      "id": "ad-42",
      "platform": "linkedin",
      "advertiser": "Example SaaS",
      "headline": "Stop losing leads after the demo",
      "text": "Automate follow-ups in minutes. Start a free 14-day trial.",
      "cta": "Start free trial",
      "firstSeen": "2026-06-01",
      "lastSeen": "2026-07-15"
    }
  ],
  "analysisGoal": "creative_testing",
  "brandContext": "B2B sales automation for companies with 20-200 employees",
  "maxAds": 25
}

For larger workflows, pass datasetId instead. Dataset input takes priority over the UI sample and does not require downloading or reshaping the source data first.

🔗 Connect your ad datasets

The Actor normalizes common fields automatically, including:

• copy: text, body, caption, description, headline, title;
• CTA: cta, ctaText, callToAction;
• identity: id, adId, libraryId, advertiser, pageName, brand;
• format: format, adFormat, creativeType, mediaType;
• provenance: platform, source, detailUrl, sourceSearchUrl;
• dates and metrics when supplied.

This makes it suitable as a decision layer after LinkedIn Ads Library, TikTok Creative Center, Meta Ads Library, or another Apify Actor.

🧭 Choose the analysis goal

🔍 Competitor research

Compare competitor hooks, offers, proof, CTAs, audiences, and recurring creative patterns.

🧪 Creative testing

Generate specific hook, proof, offer, CTA, and audience-framing variations for an experimentation backlog.

💬 Message mining

Extract repeated pains, outcomes, objections, promises, and social-proof patterns across many ads.

🩺 Campaign audit

Find unclear offers, weak CTAs, funnel mismatches, unsupported claims, missing proof, and compliance-review risks.

📦 Example output

{
  "adId": "ad-42",
  "platform": "linkedin",
  "advertiser": "Example SaaS",
  "hookType": "pain avoidance",
  "primaryHook": "Stop losing leads after the demo",
  "offer": "Free 14-day trial",
  "cta": "Start free trial",
  "funnelStage": "conversion",
  "targetAudience": ["sales teams", "revenue operations leaders"],
  "angles": ["lead loss", "CRM admin time", "speed to value"],
  "testIdeas": [
    "Test the lead-loss hook against a CRM time-saving hook",
    "Test proof-first copy before the free-trial offer"
  ],
  "evidence": ["Stop losing leads after the demo", "free 14-day trial"],
  "confidence": 0.74,
  "longevityDays": 44
}

⚙️ Input reference

Field	Description
ads	Inline ad objects for quick tests or small batches.
datasetId	An Apify dataset to process directly. Takes priority over ads.
brandContext	Product, positioning, audience, or campaign context used to make recommendations more relevant.
analysisGoal	competitor_research, creative_testing, message_mining, or campaign_audit.
maxAds	Process 1-1,000 ads. Default: 25.
includeOriginalAd	Include the complete source record for downstream joins and audits.

🛡️ Reliable batch behavior

• Structured JSON Schema keeps output consistent across runs.
• Ad copy is treated as untrusted data, never as model instructions.
• Each GPT request is retried on transient failures.
• A failed record produces an explicit low-confidence fallback instead of disappearing silently.
• Results are streamed to the default dataset in small batches.
• Source evidence is kept short and traceable.

💵 Pricing

The launch price is $3.00 per 1,000 analyzed ads on Free and Starter plans, with lower unit pricing on larger Apify plans. Platform usage is included.

The default 25-ad run has a maximum analyzed-result charge of $0.075. Apify may also display its automatic nominal Actor-start event.

🔌 Automation and API

Run the Actor from:

• the Apify Console;
• the Apify REST API or JavaScript/Python client;
• schedules and webhooks;
• Make, Zapier, n8n, or another Actor.

Rows are written to the default dataset. Aggregate hook, angle, CTA, confidence, and fallback statistics are stored in the OUTPUT key-value record.

🔐 Privacy and responsible use

• Submit only ads and data you are authorized to process.
• Do not include API keys, passwords, customer secrets, or unnecessary personal data.
• targetAudience is a creative hypothesis, not access to platform targeting settings.
• longevityDays is calculated only when valid first-seen and last-seen dates exist.
• Compliance flags are review prompts, not legal advice.
• Creative analysis does not prove campaign performance or causation.

❓ FAQ

Does this scrape ad libraries?

No. It analyzes ads you paste or supply through an Apify dataset. Pair it with an ad-library Actor when you need collection plus intelligence.

Do users need their own OpenAI key?

No. GPT processing is configured inside the Actor.

Can different platforms be mixed in one run?

Yes. Platform and source fields are normalized per record, so one dataset may contain LinkedIn, TikTok, Meta, Google, or custom ads.

Why can confidence be low?

Public ad-library cards sometimes contain only a hashtag, short caption, or missing CTA. The Actor lowers confidence rather than filling those gaps with invented facts.

What happens when one GPT request fails?

The request is retried. If it still fails, the Actor returns an explicit fallback row with low confidence so batch cardinality and source joins remain intact.