Amazon Reviews Analyzer (AI)

Turn Amazon product reviews into a report a seller can act on: which defects to fix first, what buyers love (ready to reuse in your listing), which complaints a clearer listing would stop, how each colour or size is doing, how verified buyers differ from the rest, and why people return the product. Every quote is copied word for word from a real review.

No setup and no AI account needed. Run it with no input to get a free demo report on sample reviews.

Open the full sample report: 40 fictional reviews of a demo water bottle in two colours and four languages, with themes, listing fixes, variants, buyer groups and a monthly trend.

What you get

🧭 Fix-first list	Defects ranked by how often, how badly and how recently buyers mention them. Reviews other shoppers marked helpful count more, with a rising / steady / falling trend
💬 Themes with real quotes	Top praise and complaint themes with counts, share of reviews and up to 3 verbatim quotes each, linked to the review
✍️ Fix your listing	Complaints about size, instructions or the description that a clearer product page can stop, next to what buyers love for your bullets
🎨 Issues by variant	Reviews, average stars, negative share and top complaint for each colour, size or style
✅ Verified buyers and others	How verified purchases compare with the rest, and the share of Amazon Vine reviews
↩️ Returns and refunds	How many reviews mention a return, refund or replacement, and the complaints behind them
📊 Product comparison	Side-by-side table when you analyze several products, for example yours against competitors
📈 Month by month	Review volume and average stars per month
🚨 Needs attention now	Reviews about injuries, burns, fire, choking hazards, allergic reactions or counterfeits
📄 Three outputs	A seller-ready HTML report, a short JSON summary for code and AI agents, and one dataset row per review

How to use it

1. With product URLs or ASINs: paste Amazon product URLs (any Amazon country) or 10-character ASINs and click Start. We scrape the newest reviews with Amazon Reviews Scraper on your account first, then analyze them.
2. With a dataset you already have: put the dataset ID from an Amazon Reviews Scraper or Amazon Reviews Extractor run in Dataset ID.
3. Automatically after every scrape: open your scraper (or its saved task), go to Integrations, add Amazon Reviews Analyzer (AI). The scraper's dataset is passed in automatically, so each scheduled scrape produces a fresh report.

Then open the Report tab of the run for the HTML report, or the Summary tab for JSON.

How much does it cost

You pay per result. Running the Actor itself costs you nothing extra: no compute or proxy charges.

Event	Free & Starter	Scale	Business and up
Analyzed reviews (per 1,000)	$1.50	$1.20	$1.00
Report (per run)	$0.05	$0.05	$0.05

Never charged: the demo run, star-only reviews with no text (they still count toward stars and trends), reviews we could not analyze, and runs that find no reviews.

Examples on the Free and Starter plans:

• One product with 200 reviews: $0.30 + $0.05 report = $0.35.
• 1,000 reviews: $1.50 + $0.05 = $1.55. Apify's free $5 monthly credit covers about 3,000 reviews a month.
• A brand tracking 10 products with 200 reviews each in one monthly run: $3.00 + $0.05 = $3.05 a month.

Scraping is billed separately. If you give product URLs, Amazon Reviews Scraper charges for the scrape on your account: about $6.00 per 1,000 reviews on the Free plan, less on paid plans. On Apify's Free plan it returns at most 10 reviews for 1 product per run, so for real volumes use a paid plan or give us a dataset you already have.

Free-plan runs analyze up to 1,000 reviews each. Your run's spending limit is always respected: if it is lower than the full job, we analyze what fits and tell you in the status message.

Output

JSON summary (the OUTPUT record), shortened, from the free demo run on 40 fictional reviews of a water bottle:

{
  "coverage": { "reviews": 40, "analyzed": 40, "noText": 0, "skipped": 0 },
  "overall": {
    "avgStars": 3.23,
    "topPraise": [{ "name": "Keeps drinks cold for hours", "count": 7, "share": 0.17, "quote": "there was still ice the next morning" }],
    "topComplaints": [{ "name": "Lid leaks after use", "count": 6, "share": 0.15, "quote": "After three weeks the lid leaks whenever the bottle tips over" }],
    "fixFirst": [{ "name": "Lid leaks after use", "score": 17.351, "negativeCount": 6, "shareOfNegative": 0.3, "trend": "up" }]
  },
  "variants": [
    { "productAsin": "DEMO-TRAIL-BOTTLE", "variant": "Color: Forest Green", "reviews": 20, "avgStars": 2.75, "negativeShare": 0.6, "topComplaint": "Lid leaks after use" },
    { "productAsin": "DEMO-TRAIL-BOTTLE", "variant": "Color: Slate Grey", "reviews": 20, "avgStars": 3.7, "negativeShare": 0.3, "topComplaint": "Lid leaks after use" }
  ],
  "buyerGroups": {
    "verified": { "reviews": 33, "avgStars": 3.06, "negativeShare": 0.52 },
    "unverified": { "reviews": 7, "avgStars": 4, "negativeShare": 0.14 },
    "vineShare": 0.05
  },
  "returns": { "count": 1, "share": 0.03, "topReasons": ["Lid leaks after use"] },
  "listingFixes": [{ "name": "Too wide for cup holders", "count": 3, "share": 0.07 }]
}

Dataset, one row per review: product ASIN and title, variant, verified purchase, Vine, helpful votes, country, stars, date, review title and text, analysis status, sentiment and score, aspects, phrases with verified excerpts, urgent flag and reason. Export it to Excel, CSV or JSON from the Storage tab.

How accurate is it

On a hand-labeled test set of 60 Amazon reviews of electronics, kitchenware, clothing, toys and beauty products in 7 languages (2026-10-18):

Check	Result
Overall sentiment matches the label	96.7% to 100%
Aspects found (F1)	0.88 to 0.90
Urgent reviews caught	100%, with no false alarms
Quotes that are word for word from the review	100% (enforced by code: anything else is dropped)

The AI only labels reviews. Every count, share, score and trend is computed by code from those labels, so the numbers add up.

What it does not do: it does not read review photos or videos, does not detect fake reviews, and cannot analyze star-only reviews (they still count in stars and trends). It does not draft replies, because sellers cannot reply to Amazon reviews in public. AI labels can be wrong on sarcasm or very short reviews, which is why every theme shows its quotes.

For AI agents and automations

• Free demo: call it with {} to get a full sample report and JSON summary at no cost.
• Minimal input: { "productUrls": ["https://www.amazon.com/dp/B0..."] }, { "productUrls": ["B0..."] } or { "datasetId": "<dataset id>" }.
• Read the result: the OUTPUT record is a short JSON summary designed to fit in one tool call (under 20,000 characters; with many products, per-product detail is shortened first): coverage, average stars, sentiment split, top 3 praise and complaint themes with one quote each, fix-first list, monthly trend, variants, buyer groups, returns, listing fixes and a link to the HTML report.
• Never charged: no reviews found, star-only reviews, and reviews that could not be analyzed. The status message says why.
• Works through Apify's MCP server, the Apify API and SDKs, and Make, Zapier or n8n.

FAQ

Which languages does it support? Reviews in any language, from any Amazon country. Theme names are in the language you choose (reportLanguage, default English) and report headings are in English. Quotes stay in each review's own language.

Can I compare my product with competitors? Yes. Give several product URLs or ASINs (or a dataset with several products) and the report adds a product comparison table, with a summary per product in the JSON.

Why does the report show an ASIN instead of the product name? Your dataset was scraped without product details, so its reviews carry no product title. Give product URLs instead (we switch product details on for you), or turn on product details (scrapeProductDetails) in Amazon Reviews Scraper.

Does it collect reviewer names? No. When we scrape for you, personal data collection is switched off. Reviewer names, profile links and avatars are removed from any dataset you give us before anything else happens, and never appear in the output.

Where is review text processed? By the Anthropic API (Claude). SignalSift keeps nothing after the run; your results stay in your Apify storage.

Is it legal? It analyzes publicly available reviews. You are responsible for how you use the results and for complying with the laws that apply to you, including GDPR when you handle personal data.

Something is wrong or missing? Open an issue on the Issues tab. We read every one.

Changelog

• 0.1.6 (2026-10-18): Fixed: a product given twice (as an ASIN and as its link, for example) made the run fail before anything was charged. It is now scraped, budgeted and charged once.
• 0.1.5 (2026-10-18): Catch-all themes read "Other complaints" instead of "Other other complaints".
• 0.1.4 (2026-10-18): Quotes under Fix your listing show stars, date and a link, like every other quote. The variant table says how many variants it leaves out, and the JSON summary has variantsOmitted. In the JSON summary, buyerGroups is null when every review is in one group (it compares nothing then). A run restarted by the platform waits for the scrape it already started instead of paying Amazon Reviews Scraper twice. Review titles and texts are trimmed.
• 0.1 (2026-10-18): First release. Product URLs, ASINs, datasets and integrations; themes with verified quotes; fix-first list weighted by helpful votes; listing fixes; issues by variant; verified buyers and Vine share; returns and refunds; product comparison; monthly trend; urgent flags; HTML report and JSON summary; free demo.

---

SignalSift is not affiliated with or endorsed by Google, Amazon, TikTok, Meta or YouTube. Their logos only show which platform each Actor works with.