Your Amazon reviews already contain everything you need to sell more books. You're just not reading them the right way.
I don't mean skimming your own reviews for ego boosts (or damage control). I mean systematically analyzing what readers say about books in your genre — yours and your competitors' — to understand what language they use, what they care about, and where existing books fail them.
The readers who leave reviews are doing free market research for you. They're telling you exactly what words trigger a purchase, what promises your blurb should make, and what gaps exist in your category that nobody's filling. Most authors ignore this data or read it casually. I wanted to extract it properly.
So I built a tool.
What Review Miner Does
Review Miner is a command-line tool that imports Amazon reviews and uses AI to analyze them across several dimensions:
Sentiment analysis. Not just positive/negative — it breaks down what specifically readers loved and hated. Pacing? Characters? World-building? The ending? You get a structured breakdown, not a star rating.
Ad copy generation. The best ad copy comes from your readers' own words. Review Miner extracts the most compelling phrases and patterns from positive reviews and generates ad copy variations you can test directly in Amazon Ads. Real reader language converts better than anything you'll write in a vacuum.
Gap analysis. What are readers complaining about in your competitors' books? Those complaints are your opportunities. Review Miner identifies the recurring criticisms in a category and shows you exactly where you can differentiate.
Competitor comparison. Import reviews from multiple books and compare them side by side. Who has stronger characters? Who has pacing problems? Where does your book win, and where does it lose?
Monitoring. Set up automated checks on your books. Review Miner tracks new reviews over time and sends you email digests so you're not compulsively refreshing your Amazon page.
Why This Matters More Than You Think
Most indie authors write their blurb once and forget about it. They write ad copy based on what they think sounds good. They guess at their positioning.
But your readers are literally telling you what works. When someone writes "I couldn't put it down after chapter 3," that's not just a nice review — it's a data point. When five different reviewers mention "the twist at the end," that's your hook. When a competitor's readers keep saying "great concept but the pacing dragged in the middle," that's your opening.
Review Miner turns that scattered feedback into structured data you can act on.
What You Need
Python 3.9+ and an Anthropic API key (for the AI analysis). The API costs are minimal — analyzing 100 reviews runs about $0.15.
git clone https://github.com/rxpelle/review-miner.git
cd review-miner
pip install -e .
Import reviews from a CSV export, then run the analysis:
review-miner import --file reviews.csv --title "My Book" --asin B0EXAMPLE
review-miner analyze 1
review-miner adcopy 1
review-miner gaps 1
The Honest Limitations
It's a command-line tool. No web dashboard.
It requires you to export reviews yourself (Amazon doesn't offer a public review API). You can copy-paste from your KDP reports or use a scraping tool to get competitor reviews.
The AI analysis is only as good as the input — a book with 5 reviews won't give you the same depth as one with 200. But even a handful of reviews surface patterns you'd miss reading them one at a time.
Try It
Full docs and source code on GitHub:
github.com/rxpelle/review-miner
If you find it useful, the best way to say thanks is to check out my books and leave an honest review. You'd be contributing to the very dataset that makes this tool work.