Amazon Product Monitor

Stop guessing which products to sell. Every listing gets scored for demand, competition, and profit — before you export.

config.py

---

Why This Exists

Most Amazon scrapers hand you raw data — prices, ratings, reviews — and leave you to figure out what it means. You still need to decide: Is this product worth selling? How competitive is this niche? What's the profit potential?

This Actor answers those questions. Every product gets scored before you open the spreadsheet.

You get this	Competitors give you
Demand score (0-100) + reasons why	Raw numbers. You decide.
Competition score + profit potential	Nothing. It's on you.
AI sales intelligence (free)	No AI. Or paid AI surcharges.
9 discovery modes	1-3 modes
59 output fields	~30 fields
4 export formats (JSON, CSV, JSONL, SQLite)	JSON only (sometimes)
Schedule-ready with resume	Stateless. Start over.
Zero-cost AI enrichment	$0.003-0.05/row AI fees

The bottom line: You're already paying for raw data. Getting intelligence at the same price means every dollar works twice as hard.

---

Who This Is For

• Amazon sellers — find profitable products and know when to enter a market
• Dropshippers — score products before listing and spot trends early
• Arbitrage sellers — identify price gaps with profit potential scores
• Affiliate marketers — discover trending products with high demand
• E-commerce agencies — run competitive analysis for clients at scale
• AI agents — feed structured, enriched product data into automated workflows

---

What You Get Per Product

Each product you scrape comes back with 59 structured fields — not just raw data, but actionable intelligence.

Pricing & identity — ASIN, title, brand, category, current price, list price, discount %, coupon, currency, URL, images

Ratings & demand signals — star rating, review count, bestseller rank, Prime eligibility, fulfillment type, stock status, seller info, delivery estimate

Opportunity scores (0-100, each with confidence + reasons):

• Demand — how many people want this product
• Competition — how hard it is to enter this niche
• Profit potential — what the margin opportunity looks like
• Opportunity — composite: should you sell this?

Trend & market signals — rising/steady/declining direction, seasonality, resale potential, bundle detection, anomaly detection

AI sales intelligence (free, zero API cost):

• Search intent — what customers are actually looking for
• Target customer — who buys this product
• Suggested keywords — for your listings and PPC campaigns
• Sales angle — how to market it effectively
• Competitive strengths and weaknesses
• One-line opportunity verdict

Every score works with or without an AI API key. Deterministic signals power the core scoring; AI enriches it when available.

sample_outputs/

---

One Actor, 9 Discovery Modes

{"mode": "search", "query": "wireless mouse", "maxProducts": 20}
{"mode": "asins", "asins": ["B09X7MPX8L", "B08JHCVHTY"]}
{"mode": "urls", "startUrls": ["https://www.amazon.com/dp/B09X7MPX8L"]}
{"mode": "bestsellers", "category": "electronics"}
{"mode": "movers", "category": "home"}
{"mode": "new_releases", "category": "kitchen"}
{"mode": "store", "store_url": "https://www.amazon.com/stores/page/..."}
{"mode": "brand", "query": "Sony"}
{"mode": "category", "query": "Pet Supplies"}

---

Quick Start

On Apify Cloud (recommended)

1. Open the Actor on Apify Store
2. Select a discovery mode, enter your keyword, ASINs, or URLs
3. Click Run — results appear in the dataset with all 59 fields
4. Export from the dataset to JSON, CSV, JSONL, or SQLite

Locally with Apify CLI

# Install CLI (one-time)
curl -fsSL https://apify.com/install-cli.sh | bash

# Clone and run
git clone <your-repo>
cd amazon-store-scraper
apify run  # reads local INPUT.json

Locally with Python

python3 -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
playwright install chromium
cp .env.example .env

python -m src search "wireless mouse" --max 10
python -m src bestsellers electronics --max 20
python -m src urls https://www.amazon.com/dp/B09X7MPX8L

From JSON pipe

$echo '{"mode":"search","query":"usb cable","maxProducts":5}' | python -m src

examples/

---

Output

On Apify Cloud, results go to your run's Dataset — export to JSON, CSV, JSONL, or SQLite from the Console or API. The dataset is the canonical output.

Locally, every run produces 4 files in output/:

Format	File	Best for
JSON	output/products.json	APIs, debugging, AI agents
CSV	output/products.csv	Excel, Google Sheets, pivot tables
JSONL	output/products.jsonl	Streaming, line-by-line processing
SQLite	output/products.sqlite	Persistent analysis, SQL queries

---

AI Enrichment — Zero Cost

All AI runs on OpenRouter free models only — no API tokens consumed, no surcharge.

Signal	What it tells you
Demand score	0-100 hybrid (deterministic signals + AI overlay)
Competition analysis	Review density, price pressure, brand strength
Profit potential	Price gap, margin opportunity, category health
Trend detection	Rising, steady, or declining
Seasonality	Year-round, Q4, summer, holiday
Product intelligence	Search intent, target customer, keywords, sales angle
Safety checks	Bundle detection, anomaly detection (suspicious pricing)

Model chain: google/gemma-4-31b-it:free → nvidia/nemotron-3-nano-30b-a3b:free → fallback

Disable with ENABLE_AI=false. The Actor still scores products using deterministic signals — no API key required.

---

Environment Variables

Variable	Default	What it does
OPENROUTER_API_KEY	—	Enables AI enrichment (free models)
MAX_PRODUCTS	20	Products per mode
MAX_PAGES	3	Search pages to crawl
REQUEST_DELAY	1.5s	Seconds between page visits
HEADLESS	true	Browser visibility
ENABLE_AI	true	Toggle AI enrichment
ENABLE_PROXY	false	Toggle HTTP proxy
ENABLE_RESUME	true	Persist crawl state across interruptions

AGENTS.md

---

Pricing

Tier	Results/month	Price	AI Enrichment
Free	10	$0	✅
Standard	Per 1K results	$0.75	✅

AI is always included at no extra cost. Pay per result, not per month.

---

FAQ

Do I need a paid API key for AI? No. All AI uses OpenRouter free models. No API key = AI skipped, deterministic scoring still runs. Every score works with or without AI.

What do the opportunity scores mean? All scores are 0-100. Demand = how many people want this (reviews, rating, BSR, Prime). Competition = how hard it is to enter (review density, discount depth, brand strength). Profit potential = margin opportunity (price gap, category health, brand premium). Opportunity = composite: demand × (1 − competition) × profit.

What Amazon regions are supported? Currently amazon.com (US). Multi-market (DE, UK, JP, CA, AU, FR, IT, ES) is on the roadmap.

Can I scrape 1,000+ ASINs? Yes — use bulk ASIN or URL input. Throughput is ~6-8 products/minute with sync Playwright. For large batches, batch into multiple runs with lower concurrency.

How often do CSS selectors break? Amazon changes HTML ~2-4 times per year. Each change requires a selector update (~1-2 hours). The Actor's retry + backoff handles transient issues automatically.

Is scraping Amazon legal? Public product data (prices, ratings, descriptions) is legal to scrape. See Apify's scraping legality guide.

How do I use this with Apify schedules? Create a daily or weekly schedule in Apify Console. The Actor's resume support picks up where it left off if interrupted. Structured logs are written to every run.

---

Roadmap

• 9 discovery modes
• Opportunity scoring (demand, competition, profit)
• AI intelligence (intent, keywords, sales angle)
• 4 export formats (JSON, CSV, JSONL, SQLite)
• Deterministic scoring without AI
• Multi-market (DE, UK, JP, CA, AU, etc.)
• Change detection (price/rating/review delta)
• Scheduled weekly reports
• Integration templates (n8n, Make, Zapier)

---

Documentation

• AGENTS.md
• examples/
• sample_outputs/
• CHANGELOG.md
• STORE_ASSETS.md

---

License

MIT