Amazon Bestsellers Scraper + AI: Rising Products Only

New Amazon products worth sourcing — rising, beatable, never repeated — in your inbox every week.

⚡ First 25 product opportunities free · 💵 $0.0037 per qualified product · 🤖 ≈ $0.012 with AI, tokens included, on plans with model access or your key · ⏱️ demo in seconds

Try it in 30 seconds. Click Try it — the form is pre-filled with seven sample list rows. Your first 25 qualified products are free; products that fail your filters cost nothing here. Type a niche such as kitchen to scan live — the Actor reads Amazon's public list pages itself. → What a run costs · What the AI adds · Run it weekly

⚡ At a glance

What you get	A sourcing shortlist: one row per product across Best Sellers, Movers & Shakers, New Releases and Most Wished For — score, pursue / watch / skip verdict, notes, angle, risks
You provide	A niche in a word (kitchen, pet supplies) and a marketplace (US, UK, DE, IN, CA, JP) — or your own Amazon list URLs
Output	a ranked sheet, best first — open it in Excel, CSV or Google Sheets — plus an e-mail digest of what is new each run
Typical run	300 rows scanned → 45 qualified (15 %) — (estimate), a few minutes
Cost of that run	$0.77 rules only · $1.15 with AI ($0.60 source scan + 45 × $0.0037 / $0.0092 here + ≈ $0.14 tokens) — the scan is $0 on the built-in reader
Free tier	first 25 qualified products per account
Keys / setup	Apify model access: no key · otherwise your own key — a free Gemini / Groq key = 1–3 live AI verdicts per run, paid = every product
Which plan	Free demo, built-in list reader and your own rows (list mode, template): any plan · the Store source Actor, only if Amazon blocks the reader: a plan that runs Store Actors

🎯 What this Actor does

Amazon product research from the lists you already watch: it keeps only the products an FBA or private-label seller can still win, and charges only for those.

• One row per product, not per list — the same ASIN on two lists is one row with every list and rank.
• Rising products first — rank change and new entrants against your last run: a weekly schedule alerts on what climbed, not the same top 100.
• A beatable-or-not read — a 0–100 score from rank, stars, review barrier, price band and list presence, plus flags like low_barrier.
• AI verdict with reasons — pursue / watch / skip, notes, an angle from the title and a niche, grounded in the row's own figures.
• Risks named — brand-locked, commodity and regulated titles are flagged before you order a sample.
• Niche in, lists out — type kitchen or pet supplies in niches (or just your goal) and those departments' Movers & Shakers and New Releases are scanned on the marketplace you pick (marketplace) and named in the status (OUTPUT.seedsDerived).
• No source bill on any plan — the built-in reader reads the public Best Sellers and New Releases pages itself (Movers & Shakers pages load in the browser only and are skipped, named in the result). Only if Amazon blocks every list does the Store source Actor run, capped.

🚀 How to use it

1. Click Try it — the form is pre-filled with the free demo on sample list rows; it works as is.
2. Type a niche ("Niches or departments to scan") or paste your own list URLs — it replaces the sample and starts the live scan. Leave "AI assessment" on: your first 5 products get live AI verdicts on any plan; all 25 free ones with Apify model access or a paid own key.
3. Press Start and open the dataset — qualified products land best-first; export to CSV/Excel, or wire a webhook.
4. Schedule it weekly (Actions → Schedule) with your address in notifyEmail — only new or climbing products are delivered and charged.

📦 Already scraping Amazon bestsellers? Qualify your scraper data

Paste your last scraper run's dataset ID into Dataset ID — no second scrape, you pay only for qualified products. After every scrape: the chained-run recipe.

📚 Learn more: Step-by-step tutorial — find rising Amazon products to sell

🆚 Why this instead of running the source Actor yourself?

	Fast Amazon Bestsellers Scraper (amazon-scraper/amazon-bestsellers-scraper)	Amazon Bestsellers Scraper + AI: Rising Products Only
Price	$0.002 per row + $0.00005 start · 323 users / 30 days (Store metrics, 2026-10-04)	from $0.0037 per qualified product, first 25 free, no start fee
What you pay for	every row scraped, saturated, brand-locked and duplicate ones included	only qualified products — what fails your filters costs $0 here (the source's scan is billed on your account, capped)
Same 100 usable products, all in	$1.33 for ≈ 667 raw rows — and you still sort them	$2.33 basic · $5.43 with the AI: the same $1.33 scan plus filtering — 100 ranked, explained products, not 667 rows to read
Your time per 100 products	≈ 10 h of reading rank, reviews, price and titles (6 min each — estimate)	≈ 0 h — scored, explained, risks flagged

The free Amazon Best Sellers pages win for a one-off browse; they have no export, API, schedule, rank history, scoring or dedupe.

💵 Pricing — what a run really costs

One event per qualified product, nothing else.

Event	When it is charged	Price
free-tier	your first 25 qualified products on this Actor, in any mode	$0.00
qualified-product-opportunity-basic	AI off — normalised fields, flags, rule score, rule verdict	$0.0037
qualified-product-opportunity-ai	AI on — everything above plus verdict, notes, angle, niche and risks	$0.0092

One free tier: your first 25 qualified products are free, whatever the tier. We never charge for products below minScore, failing a free filter, on your exclude list or already received — and there is no start fee. The source's scan is not ours to waive: in actor mode the source Actor bills your account $0.002 per scanned row, rejected ones included, capped by maxDiscoveryChargeUsd.

How that compares — $0.0037 is 0.99× the $0.00375 Store median for "amazon bestsellers" scrapers, but only per qualified product: one replaces ≈ 6.7 scanned rows, ≈ $0.025 per raw-row equivalent.

Worked example — 300 rows scanned → 45 qualified with AI on: source Actor $0.60 on your account + 45 × $0.0092 + ≈ $0.14 tokens → $1.15 total, about $0.026 per qualified product ($0.77 and $0.017 with AI off). When the built-in reader serves the lists the $0.60 is $0: $0.55.

At volume — 20 lists × 100 rows weekly: ≈ $4.73 a week (estimate).

AI tier, all in: ≈ $0.012 per product ($0.0092 here + ≈ $0.003 tokens: ≈ 450 in, ≈ 150 out) — for the verdict, notes, angle and risks that ≈ 6 minutes of reading by hand would produce (an estimate based on reading, not yet timed).

🤖 What the AI tier adds

The same product, AI off — facts, flags, rule score and verdict (free demo, 2026-10-04):

{"asin": "B0DEMO0001", "rank": 3, "rankChange": 14, "flags": ["rising_rank", "multi_list", "low_barrier"], "ruleScore": 82, "tier": "hot", "verdict": "pursue",
 "generatedText": "Look at Silicone Stretch Lids …: #3 on movers and shakers; 412 reviews at 4.5★; rule score 82/100."}

AI on (qualified-product-opportunity-ai, $0.0092) — judged against your goal (a real run of the AI example, free Gemini key, 2026-10-16):

{"asin": "B0DEMO0001", "fitScore": 75, "score": 75, "tier": "hot", "verdict": "pursue",
 "demandNote": "It ranks at position 3 and appears on both bestsellers and movers-and-shakers lists.",
 "competitionNote": "It features only 412 reviews while maintaining a 4.5-star rating, presenting an accessible entry barrier.",
 "angle": "Introduce an alternative color assortment or add a specialized mesh produce bag to differentiate from standard silicone stretch lids.",
 "niche": "silicone food covers", "riskFlags": [], "aiModel": "google:gemini-3.5-flash-lite"}

The AI also filters: in that run the Apple earbuds (98,214 reviews) scored below the bar, not delivered.

Try the AI on any plan: the demo shows labelled sample verdicts, your first real search gets 5 live AI verdicts on us, then bring a key (llmProvider: byok): a free Gemini key (get one) assesses a few products per run, a paid key every product.

⚙️ Input

{"niches": ["kitchen"], "marketplace": "amazon.com", "goalDescription": "Kitchen gadgets a new listing can beat",
 "maxReviews": 1500, "maxQualified": 50}

🎛️ Input options

Field	Type	Default	What it does
niches	string[]	—	Departments in plain words (kitchen, pet supplies): their Movers & Shakers and New Releases lists are scanned
marketplace	select	amazon.com	The Amazon domain the niches are read on (.co.uk, .de, .in, .ca, .co.jp)
categoryUrls	string[]	—	Your own Amazon list URLs (Best Sellers, New Releases, Movers & Shakers, Most Wished For), mixed freely
maxDiscoveryChargeUsd	number	0.5	Cap on the source Actor when Amazon blocks the built-in reader — same in the form and for an API call
goalDescription	string	—	What the AI scores each product against
minScore	0–100	50	The field that decides what you pay: below this = not delivered, not charged
maxReviews	integer	0 (off)	The review barrier: drop products with more reviews (1,500 suits a new private label)
minStars, minReviews, priceMin, priceMax, maxRank	number / integer	0 (off)	Free filters: rating, reviews, price window, top-N rank
excludeBrands, onlyNewOrRising	string[], boolean	—, false	Skip brand names; keep only new, climbing or growth-list products
enableAi	boolean	true	AI assessment on/off
maxQualified	integer	100	Hard cap on output (and on spend)

sourceMode (actor / dataset / list; {} runs the free demo) and the other options are in the Input tab; flags (rising_rank, low_barrier …) go into targetFlags — all twelve in the API reference.

📤 Output

One dataset item per qualified product — export as JSON, CSV or Excel, or stream it to a webhook.

One pinned item and the OUTPUT keys are in the API reference.

Upsert key for a CRM or Sheet: dedupeKey — marketplace + ASIN (asin:amazon.com:B0…), the cross-run memory's key: one product, one row however many lists or runs find it.

Every delivering run also writes a DIGEST (new, hot, what changed since the last run, top 10) and a client-ready REPORT page. Views: Shortlist, Sourcing sheet, Rising and new, Overview, Score breakdown.

⭐ Found it useful? A review on the Store helps others find it.

🔁 Run it weekly

1. Open the saved task and click Schedule after your first real run (its status links it). Put your address in notifyEmail — every run with new products e-mails the digest (top 10 with links and the trend vs the last run).
2. Leave dedupeAcrossRuns: true — every product you were charged for is remembered, so the next run charges only what is new. The memory is per search; give each client its own dedupeStoreName.
3. Paste a Slack, Teams, Google Chat, Discord or Telegram link into Send alerts to for the digest in your channel, or Google Sheets for the rows: start from the ready tracker sheet (dedupeKey first) and the Slack message block.
4. From the second run on, the REPORT page shows every run; historyDatasetName keeps a dated log.
5. Amazon's lists change daily, so the same lists return new entrants and climbers (week-2 yield: estimate).

Weekly cost is new products only: ≈ 10 new or climbing a week (estimate) × $0.0092 + $0.50 source ≈ $0.59 (the source part is $0 with the built-in reader).

🎯 Try it for your niche

Pick one in Start from a ready niche search at the top of the form, press Start, then schedule it.

Niche	What it looks for	Start
Kitchen gadgets	Movers & Shakers and New Releases in Kitchen, ≤ 1,500 reviews, new or rising only	preset kitchen-rising
Pet supplies	New Releases in Pet Supplies, ≤ 300 reviews, 4+ stars, $10–$50	preset pet-new-releases
Home & garden	Best Sellers + Most Wished For in Home & Garden, ≤ 3,000 reviews, no Amazon Basics / Rubbermaid / OXO	preset home-bestsellers

🔌 Integrations, automation and API

After every Amazon bestsellers scraper run (junglee/amazon-bestsellers, amazon-scraper/amazon-bestsellers-scraper): in that Actor's Integrations tab add "Run succeeded → start this Actor" with sourceMode: dataset, datasetId = resource.defaultDatasetId, enableAi: false.

Webhook, Google Sheets and Slack from the Integrations tab, Airtable / Notion / HubSpot through a Make or Zapier "create or update" step (upsert on dedupeKey; fields verdict, niche, angle, priceValue, reviewsCount, rank), and MCP for AI agents — every input is a plain typed field. The free first call is {} (the demo) or {"sourceMode": "list", "itemsList": [...]}.

Minimal inputs: search {"niches": ["kitchen"]} · URL {"categoryUrls": ["https://www.amazon.com/gp/movers-and-shakers/kitchen"]} · any plan, your own rows {"sourceMode": "list", "itemsList": [{"title": "…", "asin": "B0…"}]}.

👥 Who is it for?

You are…	You run it to…	Start with
Amazon FBA / private-label seller	find rising, beatable products in your category	preset kitchen-rising
Product researcher or sourcing agent	hand a client a shortlist with reasons	Movers & Shakers + New Releases URLs, onlyNewOrRising, the Sourcing sheet
E-commerce store owner	learn when something new enters your shelf	one URL per shelf, weekly, notifyEmail
Dropshipping / wholesale buyer	skip brand-locked and regulated listings	excludeBrands

🧠 How the AI works

• Grounded, not generative. The model sees one product's list-page fields and returns a typed assessment. It is told it knows no sales, revenue or margin; a sentence quoting a number that is not in the row, or a sales claim, is dropped and replaced by a rule-built note. A risk flag stays only if its quote is in the title.
• Your goal, not a generic score. The fit score is against your goalDescription: pursue = hot (≥ 70), watch = warm (45–69), skip = cold (no angle, no text).
• Tokens only where they matter. The best rule-scored candidates get the AI pass, best first; no key with Apify model access (anthropic/claude-haiku-4.5), else your own (a Gemini key runs google:gemini-3.5-flash-lite). If the model cannot be reached, products arrive rule-scored.

🔒 Data, compliance and limits

• Only what Amazon publishes on public list pages (title, rank, price, rating, reviews, category) — no login, no seller or reviewer data.
• A list page shows 50 products (two pages = top 100); a missing rating or price is null, never guessed; currency sits next to every price.
• The score (demand 40 · growth 25 · monetisation 20 · quality 15) ranks what to look at first — not sales data. Keep your use within Amazon's terms and applicable law (GDPR, CCPA).

❓ FAQ

How much will one run cost me? A typical run — 300 rows scanned, 45 qualified — is in the worked example above; the built-in reader makes the source scan $0.

What will my first real search cost? ≈ 30 qualified from 200 rows — the first 25 are free here, so only ≈ 5 are paid; the source Actor's bill applies only when Amazon blocks the reader (capped by maxDiscoveryChargeUsd).

How do I find rising Amazon products? Type your niche, set maxReviews at the barrier you can beat, read the verdicts and risk flags first, and schedule it.

Is this an Amazon product research tool? For the list-page half: demand momentum, review barrier, price band and risks per ASIN — it does not estimate sales or revenue.

Is this an Amazon FBA product finder, or a Helium 10 alternative? For the first filter, yes: it finds rising, beatable products from the public lists and charges per qualified product instead of a monthly seat. It does not estimate monthly sales or margin, which seller tools like Helium 10 do — use it to build the shortlist, then check the few survivors there.

Can I track Amazon Movers and Shakers weekly? Yes — add your category's Movers & Shakers URL, schedule the run, set onlyNewOrRising; each digest lists what entered or climbed.

🧩 More Actors from the same developer

Same pay-per-qualified-result model — next step after your shortlist: check what the winners resell for, or how rivals advertise them.

• eBay Sold Listings Comps AI — Clean eBay sold comps and a max buy price for your exact item, every week — the resale check for what you found here.
• Facebook Ad Library Winning Ads Finder — The Facebook ads in your product's niche that have run longest, hook and offer broken down.
• Trending Keywords for Your Niche — Your niche's new trending keywords and Black Friday peaks, in your inbox every week.
• Website Change Monitor — Every real change on your competitors' pages, explained, in your inbox every week.
• 1688 Supplier Finder — New trusted 1688 factories for your product — costed and translated, in your inbox every week.

🆘 Support

Something missing or wrong? Open an issue on the Actor's page.

📝 Changelog

• 0.1.6 (2026-10-16) — One "Send alerts to" box takes your Slack, Discord, Teams, Google Chat, Telegram or webhook link, and the form is shorter: technical options now sit out of the way; A product you already paid for under another search or filter set is still delivered there but never charged twice, and webhook posts can be signed so your endpoint knows they came from this Actor; If your plan cannot run the backup Amazon scraper, the run now ends calmly with a clear next step instead of failing.
• 0.1.5 (2026-10-05) — the live scan reads Amazon's Best Sellers and New Releases pages itself, so it works on any plan with no source bill (Movers & Shakers pages are skipped and named in the result); type a niche and pick a marketplace; OUTPUT.outcome says ok, empty or source_failed; the hot / warm / cold field is tier (label keeps the same value for one more version); webhook events are product.qualified / products.qualified; a Slack message block ships with the weekly recipe.
• 0.1.2 (2026-10-05) — pricing now applies on the Apify Store: each qualified product is billed at $0.0037 (rules only) or $0.0092 (AI-assessed), first 25 free.
• 0.1.1 (2026-10-05) — initial release: Amazon list URLs (or just your goal) in, one row per ASIN with rank, rank change, price, a score and a pursue / watch / skip verdict.