ProductPulse – AI Product Data Extraction & Copywriting

The fastest way to enrich your product catalog. Extract GTINs, images, and product attributes – plus AI-generated descriptions like a professional copywriter.

• 📦 Batch Processing – Up to 100 products in a single run
• ✍️ AI Copywriter – Professional product descriptions in 6 styles, any language
• 🖼️ Auto Image Discovery – Official product photos extracted automatically
• 🧠 AI-Powered Intelligence – Fresh data from web sources, not stale databases
• 📊 Confidence Scoring – Know exactly what to trust vs. review

---

⚡ Power Features

Batch Processing (Up to 100 Products)

Process entire product lists in a single API call. Perfect for catalog enrichment at scale.

Batch Size	Processing Time	Cost
10 products	~3-5 minutes	$0.90
50 products	~5-8 minutes	$4.50
100 products	~8-12 minutes	$9.00

Benefits:

• ⚡ Parallel processing architecture
• 📊 Partial success handling (one failure doesn't stop the batch)
• 💰 Same $0.09/product pricing at any volume

Auto Image Discovery

ProductPulse automatically finds official product images from authoritative web sources:

"image_urls": [
  "https://static.nike.com/a/images/t_PDP_1728_v1/f_auto/air-max-90.jpg",
  "https://images.footlocker.com/nike-air-max-90-white.png"
]

• 🔍 Discovers images from manufacturer sites, retailers, and official sources
• 📸 Returns multiple image URLs when available
• ✅ Prioritizes high-resolution official product photos

---

🎯 The Problem We Solve

Manual product research is a time sink:

Task	Manual Effort	With ProductPulse
Find GTIN/EAN code	5-15 minutes	⚡ Automatic
Write product description	10-20 minutes	⚡ AI-generated in 5 languages
Categorize product	3-5 minutes	⚡ Auto-classified
Extract specifications	10-15 minutes	⚡ All attributes at once
Total per product	30-55 minutes	⏱️ 60-120 seconds

You have 500 new products from a supplier with only names and brands? That's 250+ hours of manual work. Or 15 hours with ProductPulse.

---

🔄 How Does AI Product Data Extraction Work?

┌─────────────────────────────────────────────────────────────────────┐
│                                                                     │
│   📝 INPUT              🤖 AI INTELLIGENCE            📊 OUTPUT    │
│                                                                     │
│  ┌───────────────┐     ┌─────────────────────┐    ┌─────────────┐ │
│  │ Product Name  │     │                     │    │ GTIN/EAN    │ │
│  │ Brand         │────▶│   Web Intelligence  │───▶│ Categories  │ │
│  │ Language      │     │   + AI Analysis     │    │ Images      │ │
│  └───────────────┘     │   + Verification    │    │ Descriptions│ │
│                        │                     │    │ + 20 more   │ │
│                        └─────────────────────┘    └─────────────┘ │
│                                                                     │
└─────────────────────────────────────────────────────────────────────┘

Three simple steps:

1. Send product name + brand → Minimal input required
2. AI gathers web intelligence → Multi-source verification from authoritative sources
3. Get structured data → JSON with confidence scores for every attribute

---

❓ What Product Data Can You Extract Automatically?

Category	Attributes	Confidence Scored
🏷️ Identification	GTIN/EAN barcode numbers	✅ Yes
📦 Basic Data	Brand, product name, category	✅ Yes
🎨 Appearance	Primary color, material	✅ Yes
📐 Dimensions	Size information	✅ Yes
🖼️ Images	Product image URLs from official sources	✅ Yes
✍️ AI Copywriting	Professional product descriptions in any language	✅ Yes
📊 Quality Metrics	Completeness %, review recommendations	✅ Yes
🔍 Extended Attributes	AI-discovered bonus attributes (technical specs, features)	✅ Yes

✍️ AI Product Copywriter

ProductPulse doesn't just extract data – it writes compelling product descriptions like a professional copywriter:

• 6 Writing Styles: Professional, Casual, Technical, Marketing, Luxurious, Minimalist
• Any Language: Generate descriptions in German, English, French, Spanish, Italian, or any other language
• Product Highlights: Emphasize specific selling points in the generated text
• Brand Voice: Consistent tone across your entire catalog

Example output:

"The Apple iPhone 15 Pro Max represents the pinnacle of smartphone engineering.
It features a robust yet lightweight aerospace-grade titanium design with a
textured matte-glass back. The device is powered by the A17 Pro chip, delivering
industry-leading performance and efficiency."

No more generic descriptions. No more copy-paste from competitors. Every product gets unique, SEO-friendly content.

Every attribute includes a confidence score (0-1):

• 0.85+ → Auto-accept into your PIM
• 0.70-0.85 → Quick human review recommended
• Below 0.70 → Manual verification needed

---

💡 Real-World Use Cases

📋 Scenario 1: PIM Onboarding Nightmare

"We just signed a new supplier. They sent us an Excel with 2,000 product names and nothing else. Launch is in 3 weeks."

Solution: Batch process all products with ProductPulse. Get GTINs, categories, and descriptions for every item. Import directly to your PIM.

Result: 3 weeks of manual work → 3 days of automated processing.

---

🛒 Scenario 2: E-commerce Catalog Gaps

"Our product listings are missing categories, colors, and materials. Google Shopping is rejecting our feed."

Solution: Run existing products through ProductPulse to fill missing attributes. Confidence scores tell you what's reliable.

Result: Shopping feed acceptance rate: 45% → 94%.

---

🏭 Scenario 3: Supplier Data Chaos

"Every supplier uses different naming conventions. We need standardized data for our ERP."

Solution: Use ProductPulse to normalize all supplier data to your schema. AI extracts consistent attributes regardless of input format.

Result: One unified product database from 15 different supplier formats.

---

🚀 Quick Start – How to Extract Product Data

Input (Minimal Required)

{
  "productName": "Air Max 90",
  "brand": "Nike",
  "language": "de"
}

Field	Type	Required	Description
productName	string	✅ Yes	The product name to research
brand	string	✅ Yes	The product brand or manufacturer
language	string	❌ No	Output language: de, en, fr, es, it (default: de)

Output (The Money Shot)

{
  "products_attributes": {
    "gtin": "0195239925344",
    "ean": "195239925344",
    "brand": "Nike",
    "product_name": "Nike Air Max 90",
    "category": "clothing",
    "primary_color": "White/Black",
    "material": "leather",
    "size": "EU 36-49",
    "product_description": "Der Nike Air Max 90 ist ein legendärer Sneaker...",
    "image_urls": [
      "https://static.nike.com/a/images/t_PDP_1728_v1/f_auto/air-max-90.jpg"
    ],
    "extended_attributes": {
      "features_cushioning": "Air Max visible air unit",
      "features_closure": "Lace-up"
    }
  },
  "confidence": 0.92,
  "sources": [
    { "url": "https://www.nike.com/...", "title": "Nike Official" }
  ],
  "metadata": {
    "processingTimeMs": 45000,
    "textStyleUsed": "Professional"
  },
  "quality": {
    "completeness": 0.85,
    "manualReviewRecommended": false,
    "lowConfidenceAttributes": []
  }
}

---

📦 Batch Processing

Process multiple products in a single run for maximum efficiency.

Batch Input

{
  "products": [
    { "productName": "Air Max 90", "brand": "Nike" },
    { "productName": "iPhone 15 Pro", "brand": "Apple" },
    { "productName": "Dyson V15", "brand": "Dyson", "gtin": "5025155080511" }
  ],
  "language": "de",
  "ai_text_style": "Professional"
}

Field	Type	Required	Description
products	array	✅ Yes	Array of products (max 100)
products[].productName	string	✅ Yes	Product name to research
products[].brand	string	✅ Yes	Product brand
products[].gtin	string	❌ No	Optional GTIN for this product
products[].ai_product_highlight	string	❌ No	Optional highlight for this product
language	string	❌ No	Shared language for all products
ai_text_style	string	❌ No	Shared text style for all products

Batch Output

Each successful product is pushed as an individual dataset item. The key-value store contains a summary:

{
  "summary": {
    "total": 10,
    "successful": 9,
    "failed": 1,
    "successRate": 90
  },
  "results": [
    { "index": 0, "productName": "Air Max 90", "brand": "Nike", "success": true, "confidence": 0.92 },
    { "index": 1, "productName": "iPhone 15 Pro", "brand": "Apple", "success": true, "confidence": 0.95 },
    { "index": 2, "productName": "Unknown Product", "brand": "Unknown", "success": false, "error": "Timeout" }
  ]
}

Batch Pricing

Same $0.09 per product – no volume penalty or discount.

Batch Size	Price	Processing Time
10	$0.90	~3-5 minutes
50	$4.50	~5-8 minutes
100	$9.00	~8-12 minutes

Batch Best Practices

1. Optimal batch size: 10-50 products for best efficiency
2. Error handling: Check summary.failed count and individual results
3. Progress monitoring: Watch Actor status message for real-time progress
4. Rate limits: Larger batches may hit API rate limits; results still complete via retry

---

⚡ API Integration Examples

cURL

curl -X POST "https://api.apify.com/v2/acts/YOUR_USERNAME~productpulse/runs" \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"productName": "Air Max 90", "brand": "Nike", "language": "de"}'

Python

from apify_client import ApifyClient

client = ApifyClient("YOUR_API_TOKEN")
run = client.actor("YOUR_USERNAME/productpulse").call(
    run_input={
        "productName": "Air Max 90",
        "brand": "Nike",
        "language": "de"
    }
)

# Get results
items = client.dataset(run["defaultDatasetId"]).list_items().items
print(items[0])

JavaScript / Node.js

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: 'YOUR_API_TOKEN' });
const run = await client.actor('YOUR_USERNAME/productpulse').call({
  productName: 'Air Max 90',
  brand: 'Nike',
  language: 'de',
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items[0]);

Batch Processing (Python)

from apify_client import ApifyClient

client = ApifyClient("YOUR_API_TOKEN")
run = client.actor("YOUR_USERNAME/productpulse").call(
    run_input={
        "products": [
            {"productName": "Air Max 90", "brand": "Nike"},
            {"productName": "iPhone 15 Pro", "brand": "Apple"},
            {"productName": "Dyson V15", "brand": "Dyson"}
        ],
        "language": "de"
    }
)

# Get all results
items = client.dataset(run["defaultDatasetId"]).list_items().items
print(f"Processed {len(items)} products")

# Check summary in key-value store
kvStore = client.key_value_store(run["defaultKeyValueStoreId"])
summary = kvStore.get_record("OUTPUT")
print(f"Success rate: {summary['value']['summary']['successRate']}%")

Zapier / Make / n8n

Use the official Apify integration to trigger ProductPulse from your automation workflows. Results flow automatically to subsequent steps.

---

📊 How Does ProductPulse Compare to Alternatives?

Feature	ProductPulse	Manual Research	GTIN Databases	Traditional Scrapers
Time per product	60-120 sec	30-55 min	Instant lookup	5-30 sec
GTIN/EAN included	✅ Yes	✅ Manual	✅ Yes	❌ Rarely
Product images	✅ Auto-discovered	✅ Manual	❌ No	⚠️ Sometimes
Product descriptions	✅ AI-generated	✅ Manual	❌ No	❌ No
Multi-language	✅ 5 languages	❌ One at a time	❌ No	❌ No
Confidence scores	✅ Yes	❌ No	❌ No	❌ No
Legal/Ethical	✅ Legitimate APIs	✅ Yes	✅ Yes	⚠️ ToS risks
Adapts to changes	✅ Real-time web	❌ No	⚠️ Database lag	⚠️ Breaks often
Cost per product	$0.09	~$5-15 (labor)	$0.01-0.05	$0.001-0.01
Data completeness	85%+	Variable	IDs only	Variable

Why AI-powered extraction beats traditional approaches:

• 🔄 No maintenance – AI adapts to website changes
• ✅ No ToS violations – Uses legitimate API-based data access
• 📊 Quality signals – Know what to trust with confidence scores
• 🌍 Multi-source verification – Cross-references multiple sources

---

💰 How Much Does AI Product Data Extraction Cost?

Pricing: $0.09 per Product (All-Inclusive)

No hidden fees. No separate compute charges. You pay exactly $0.09 per successful product extraction.

What's Included in $0.09
AI processing & analysis	✅ Included
Web intelligence gathering	✅ Included
Apify compute resources	✅ Included
Storage & dataset operations	✅ Included
Proxy & network costs	✅ Included

What You Get

Feature	Included
Full attribute extraction (GTIN, category, colors, etc.)	✅
Product image URLs (auto-discovered)	✅
AI-generated descriptions	✅
5 output languages	✅
6 text style options	✅
Custom attributes	✅
Extended AI-discovered attributes	✅
Confidence scores	✅
Source URLs	✅
Quality metrics	✅

Volume Examples

Products	Total Cost	Cost Per Product	Time Saved vs Manual
1	$0.09	$0.09	~35 minutes
10	$0.90	$0.09	~6 hours
100	$9.00	$0.09	~60 hours
1,000	$90.00	$0.09	~600 hours
10,000	$900.00	$0.09	~6,000 hours

Pricing Transparency

ProductPulse uses Apify's Pay-Per-Event (PPE) pricing model:

• You pay: $0.09 per product extraction event
• No additional charges: Compute units, storage, and AI costs are absorbed by us
• Predictable costs: Know exactly what you'll pay before running
• Failed extractions: Not charged (only successful results trigger billing)

💡 New Apify accounts include free credits to test the Actor.

---

🌍 Supported Languages

Code	Language	Description Quality
de	German	⭐⭐⭐⭐⭐ Native quality
en	English	⭐⭐⭐⭐⭐ Native quality
fr	French	⭐⭐⭐⭐ High quality
es	Spanish	⭐⭐⭐⭐ High quality
it	Italian	⭐⭐⭐⭐ High quality

---

🔧 Troubleshooting

"Product not found" or low completeness score

Causes:

• Very niche or new product with limited web presence
• Misspelled product name or brand
• Regional product not available internationally

Solutions:

1. Verify the exact product name from manufacturer website
2. Try alternative product name variations
3. Check if product exists in major retailers

Low GTIN confidence score

Causes:

• Multiple product variants (colors/sizes) with different GTINs
• Product is very new or discontinued
• Limited authoritative sources

Solutions:

1. Include variant details in product name (e.g., "Air Max 90 White Size 42")
2. Use the extracted sources to manually verify
3. Cross-reference with manufacturer database

Processing timeout

Causes:

• Complex product requiring extensive research
• Temporary API delays

Solutions:

1. Retry the extraction
2. If persistent, contact support

---

❓ Frequently Asked Questions

Why does extraction take 60-120 seconds?

Unlike simple database lookups or scrapers, ProductPulse performs deep AI research:

1. Gathers data from multiple authoritative sources
2. Analyzes web pages and product information
3. Cross-references information across sources
4. Generates confidence scores
5. Creates multi-language descriptions

This delivers verified, high-quality data – not just cached information.

Is this legal? What about website terms of service?

Yes, ProductPulse is fully legal and ethical:

• ✅ No scraping – Uses legitimate API-based data access
• ✅ No ToS violations – AI researches publicly available information
• ✅ GDPR compliant – No personal data extracted or stored
• ✅ Ethical AI – Following industry best practices for responsible AI usage

How accurate are GTIN/EAN codes?

Product Type	Expected Accuracy
Popular brands (Nike, Apple, Samsung)	~95%
Mid-tier brands	~85%
Niche/local products	~70%
Very new products	~60%

The confidence score tells you how reliable each GTIN is. Trust high scores, verify low ones.

Can I use custom attributes?

Yes! Define your own attribute schema in the input. The AI will extract exactly what you need. Examples:

• Technical specifications (voltage, wattage, capacity)
• Fashion attributes (sleeve length, neckline, fit)
• Food attributes (ingredients, allergens, nutrition)

What if the AI gets something wrong?

Every attribute includes a confidence score. Low confidence = automatic review flag. Use the quality.manualReviewRecommended field to identify items needing human verification.

Can I process multiple products at once?

Yes! ProductPulse supports batch processing up to 100 products per run.

{
  "products": [
    { "productName": "Air Max 90", "brand": "Nike" },
    { "productName": "iPhone 15 Pro", "brand": "Apple" },
    { "productName": "Dyson V15", "brand": "Dyson" }
  ],
  "language": "de"
}

Batch benefits:

• ⚡ Parallel processing (~3-8 minutes for 50 products)
• 💰 Same $0.09/product pricing (no volume penalty)
• 📊 Partial success handling (one failure doesn't stop the batch)
• 📈 Progress tracking via status messages

---

📋 Current Limitations & Roadmap

Known Limitations

Limitation	Workaround
60-120s processing time per product	Required for quality research; batch mode processes in parallel
Max 100 products per batch	Run multiple batches for larger datasets
GTIN accuracy varies	Use confidence scores to filter

Upcoming Features

• 📦 Batch processing → Shipped in v1.3.0! Up to 100 products per run
• 🖼️ Image-based product extraction → Shipped in v1.2.1! image_urls now included
• 🔔 Webhook notifications
• 📊 Custom attribute schemas
• 🔗 Direct PIM integrations

---

🆘 Support

Having issues? Need a feature?

• 📝 Issues tab on this Actor page – Bug reports, feature requests
• 💬 Apify Discord – Community support
• 📧 Contact developer – Through Apify platform

---

📝 Changelog

v1.3.0 (January 2026)

• 📦 Batch processing: Process up to 100 products in a single run
• ⚡ Parallel processing architecture (~3-8 minutes for 50 products)
• 📊 Partial success handling with detailed error reporting
• 📈 Real-time progress tracking via status messages
• 🔄 Backwards compatible (single product mode unchanged)

v1.2.1 (January 2026)

• 🖼️ Image URL discovery: Automatically finds product images from web sources
• 📸 Returns image_urls array with official product photos

v1.2.0 (January 2026)

• 💰 Pay-per-event pricing: $0.09 per product extraction
• 🔗 Apify billing system integration
• 📊 Transparent, predictable pricing

v1.1.0 (January 2026)

• 🔍 Extended attributes (AI-discovered bonus data)
• ⚙️ Custom attribute extraction
• ✍️ 6 text style options for descriptions
• ⭐ Product highlights feature
• 📦 New products_attributes output structure

v1.0.0 (January 2026)

• 🚀 Initial release
• 🤖 AI-powered product extraction engine
• 🌍 5 output languages
• 📊 Confidence scoring per attribute
• ✅ Quality assessment metrics

---

Built by 1TM on Apify
AI-powered product data extraction at scale