Mine Amazon review sentiment (no ML required)
Turn thousands of reviews into actionable complaint and praise themes with the open-source amazon-review-analyzer and the Pangolinfo MCP.
🔥 Voice of Customer: Knowing what buyers praise or complain about is gold for product improvement and listing copy. This tool extracts those themes automatically.
1. Fork and clone
git clone https://github.com/pangolinfoapi/amazon-review-analyzer cd amazon-review-analyzer cp reviews.example.json reviews.json
2. Pick the ASINs to analyze
Edit reviews.json with the products you want to understand:
[
{ "label": "my product", "asin": "B0XXXXXXXX" }
]
3. Add your Pangolinfo token
Get a token from pangolinfo.com and set it as the PANGOLIN_TOKEN repository secret. Reviews are fetched through the Pangolinfo MCP get_amazon_reviews tool.
4. Run the analysis
export PANGOLIN_TOKEN=eyJ...your_jwt python3 amazon_review_analyzer.py run python3 amazon_review_analyzer.py report
The report shows rating distribution, positive/neutral/negative counts, and the top complaint and praise keywords per ASIN.
5. Schedule it
The included GitHub Actions workflow re-analyzes on a schedule and commits the stats, so you can track sentiment drift after a product change.
Next steps
- Combine with the Keyword Rank Tracker to see if ranking changes move sentiment.
- For competitor monitoring, check the official clawdbot-competitor-monitor (★3).
- Need raw review data at scale? See the pangolinfo-amazon-scraper SDK.