AI ASIN Evaluation — A sourcing verdict in one paste.
Drop an ASIN. Get a structured verdict — buybox forecast, seller mix, brand-control posture, margin model, and the three specific risks worth watching this week. Grounded in our daily observation data, not in vibes.
Real data first, language model second.
Most "AI sourcing tools" send your ASIN to a large language model and ask it to make up a verdict. The model has no visibility into the listing's actual seller history, no read on the brand's enforcement record, no sense of how the buybox has moved this quarter. The output sounds confident and is mostly fiction.
Our evaluation pipeline inverts that. Step one is a structured retrieval — we pull the listing's full seller history, buybox win record, price-history curve, and the brand's enforcement footprint from our daily monitoring data. Step two is a deterministic scoring pass — the same six-dimension Sourcing Score that powers the picks list. Step three uses the model only to translate that structured verdict into a readable summary, with citations back to the underlying numbers.
The result is an evaluation you can argue with. Every claim in the summary is traceable to the row it came from. If the model says “buybox volatility has tripled in the last 90 days,” you can click through to the buybox-history table and see the data. No hallucinated seller counts, no invented margin estimates.
- Composite Sourcing Score with per-dimension breakdown.
- Buybox forecast — likely winner mix over the next 30 days, given current seller behavior.
- Seller landscape summary — who's on the listing, how long they've held buybox share, and whether new sellers are joining or leaving.
- Brand-control posture — MAP enforcement signals, gating posture, takedown frequency.
- Margin model — estimated post-fee margin at the current buybox price for typical wholesale cost ranges.
- Top three risks to watch this week, with the data points behind each call.
- Walmart counterpart when one exists — buybox status, price spread, opportunity flag.
A serious sourcer can pull most of these signals manually with enough time and the right tool stack — Keepa for price history, a separate scrape for seller counts, a third tool for brand enforcement, your own margin spreadsheet. Twenty minutes per ASIN, easily. On a shortlist of fifty, that's your whole afternoon.
The ASIN evaluator collapses that into a 30-second paste. The data isn't new — it's the same numbers you'd collect by hand — but the synthesis is. You can rip through fifty ASINs in twenty minutes, killing the obvious nos, flagging the obvious yeses, and saving the genuinely ambiguous middle for human deliberation.
- Resellers triaging supplier catalogs — paste in a wholesale list, kill the obvious losers, surface the keepers.
- Sourcing teams with daily intake queues — first-pass review, then human follow-up on the ambiguous calls.
- Brand owners doing competitive recon — evaluate ASINs in your category, see where you stack up and which competitors are strongest.
- M&A diligence — score every ASIN in a target catalog and produce a quality-of-revenue read.
| Plan | What you get |
|---|---|
| Scout (free) | 3 ASIN evaluations per month |
| Sourcer | 100 evaluations per month |
| Brand Owner | 500 evaluations per month, batch input |
| Portfolio | Unlimited, API endpoint, custom rubric |
See the full pricing breakdown for limits and add-ons.
Which AI model is behind the summary?
Currently Anthropic Claude. The model is only used for the writeup; the underlying scoring is deterministic and runs against our database, not against the model.
Will it hallucinate seller counts or buybox stats?
No. The summary is constrained to cite values we've already retrieved from the database. The model can't make up numbers because it isn't the source of any number — every figure is hyperlinked back to the row it came from.
Does it work on brand-gated ASINs?
Yes — the evaluation runs regardless of whether you can sell the brand. The verdict will flag gating as a risk; you decide whether to pursue ungating.
Can I evaluate Walmart ASINs directly?
The primary input is an Amazon ASIN, but the verdict will surface the Walmart counterpart automatically when one exists. Native Walmart-only evaluation is on the roadmap.
Is there an API?
Yes, on the Portfolio tier. POST an ASIN, get back a structured JSON verdict you can pipe into your own buying workflow.
Paste an ASIN. Get a verdict.
Three free evaluations per month on the Scout tier. Paste, read, decide.