Data Sources & Subprocessors
Where LumaiScope's data comes from, who processes it, and how we keep it honest.
Last updated: 16 June 2026
Where our data comes from
LumaiScope is a research platform, not a guessing machine. Almost everything you see in the product is grounded in real, measured data pulled from named sources. We'd rather show you a measured number than a confident-sounding one we made up.
Lumai AI sits on top of that data to narrate it, summarize it, and turn it into a judgment you can act on. The key thing to understand: our AI reads the measured numbers, it does not invent them. When a figure is an estimate rather than a measurement, we label it as one.
This page lays out every source we read product data from, every third party that processes data on our behalf, and how we treat the Seller Central data you choose to connect.
Where product data comes from
The market data behind your validations, searches, heatmaps, and forecasts comes from a small set of dedicated providers:
- Oxylabs and Rainforest API — public Amazon catalogue data: product details, search results, pricing, Best Sellers Rank (BSR), reviews, and seller-landscape signals.
- DataForSEO — Amazon keyword data: search volume, cost-per-click (CPC), and seasonality trends that power Keyword Research, the Demand Heatmap, and Forecasts.
- Reddit API — public social-trend signals used in Social Trends to surface what sellers and shoppers are talking about.
- Amazon Selling Partner API (SP-API) — your own Seller Central data: inventory, orders, financials, and PPC. This is accessed only with your explicit OAuth authorization, and used only to power your own dashboards (Performance, Monitoring, Reimbursements).
Subprocessors
To run LumaiScope we rely on a short list of trusted third-party services. Each one processes data only for the specific purpose listed below:
- Clerk — authentication and account/session management.
- Stripe — payment processing and subscription billing.
- Anthropic — AI processing of the product data you submit. This happens in real time to generate verdicts, listings, and analysis. Your data is not used to train AI models.
- Oxylabs and Rainforest API — public Amazon product, pricing, and BSR data.
- DataForSEO — Amazon keyword search-volume, CPC, and seasonality data.
- Reddit API — public social-trend signals.
- Resend — transactional and account email delivery.
- Sentry — application error monitoring and diagnostics.
- Amazon Selling Partner API — your authorized Seller Central data, used only to power your dashboards.
How we use Amazon Selling Partner data
When you connect Amazon Seller Central through OAuth, you're authorizing LumaiScope to read your seller data for one reason: to power the features you're using — your P&L, PPC analytics, inventory and order tracking, reimbursement scanning, and the rest.
- We use your Selling Partner data only to provide the features you've enabled.
- We never sell your Seller Central data, and we never share it for advertising.
- You can disconnect at any time, and you can request deletion of your stored seller data whenever you want.
- Our handling of this data is consistent with Amazon's Acceptable Use and Data Protection Policy.
Data accuracy & honesty
Not all numbers are the same, so we're careful about how we present them:
- Measured — pulled directly from a source (e.g. a live price, a BSR, a keyword search volume). We show it as-is.
- Estimated — modeled from measured inputs (e.g. monthly sales derived from BSR). We label estimates clearly so you know they're a calculation, not a reading.
- AI judgment — Lumai AI's read on the measured and estimated data (e.g. a GO / Watch / Skip verdict or a Deal Score). It's a judgment to weigh, not a guarantee.
We don't promise guaranteed profits or rankings, and we don't dress up an estimate as a fact. Markets move, listings change, and Amazon updates its data — so treat every verdict as one strong input into your decision, not the final word.
Questions & related policies
For more on how we handle your data and your rights over it, see our privacy and GDPR policies — or reach out and a human will help.