Amazon Review Scraper

A Claude Code skill that scrapes Amazon product reviews via woot.com's public AJAX API.

No API key. No login. No browser automation. Pure Python 3 stdlib.

Install as Claude Code Skill

# Install via npx (npm)
claude install npx:amazon-review-scraper

# Or install via GitHub
claude install github:mrlong0129/amazon-review-scraper

Usage

Once installed, just ask Claude naturally:

"Get all reviews for B0BLCBRBVZ"
"Scrape the bad reviews for B0DPHLWNBG"
"Analyze the reviews for this ASIN: B0BN7M8CH4"
"Get recent 3 months reviews for B0xxx"
"Get reviews from UK site"

Claude will automatically:

1. Confirm the marketplace (US / GB / DE / JP / etc.)
2. Scrape reviews using the optimal mode
3. Apply any filters (date, star rating)
4. Generate a full review data document and a summary analysis

Standalone CLI Usage

npx amazon-review-scraper B0BLCBRBVZ
npx amazon-review-scraper B0BLCBRBVZ --mode basic -o reviews.json
npx amazon-review-scraper B0BLCBRBVZ --summary

How It Works

Dual-Source Collection

| Source | Marketplaces | Max Reviews | Variant Attributes | |--------|-------------|-------------|-------------------| | Woot Scraper | US only | Unlimited (max mode) | No | | Sorftime MCP | 14 marketplaces | 100 / ~1 year | Yes |

For US marketplace, both sources are used with automatic dedup merge.

Scrape Modes (Woot)

| Mode | Max Reviews | Speed | Best For | |------|-------------|-------|----------| | basic | 100 | Fast | Quick preview | | full | 500 | Medium | Most products | | max | ~500-700 | Slower | Maximum extraction (default) |

Coverage

| Product Size | Expected Coverage | |-------------|-------------------| | < 100 written reviews | 100% | | 100 - 500 | ~95-100% | | 500 - 1000 | ~85-95% | | > 1000 | ~70-85% |

Note: Amazon "ratings" include star-only ratings (no text). This tool retrieves written reviews only — typically 10-15% of total ratings.

Requirements

• Python 3.6+
• No external dependencies (uses only stdlib: urllib, json, argparse)

License

MIT