How TwelveTabs works
Step 1
We research the product online.
We learn all the product differences, then ask you questions to find the perfect match.
Step 2
We read the Amazon listings & reviews.
After analyzing over 100+ listings, we read all reviews from the best matching products.
Step 3
We choose a winner with top alternatives.
We rank them based on the rating & number of reviews, amount bought last month, price, and best match.
It is free, no account is required for up to three searches a day, and it takes about a minute.
FAQs
Is it free?
Yes. There is no charge, no subscription and no trial that ends. Up to three searches a day are free without logging in, fewer when several people share one internet connection; after that, confirm your email address to keep searching. Logging in also keeps your preferences and search history across devices.
How does it pick a winner?
It turns your question into an Amazon search, then drops the sponsored placements outright, along with any row it cannot build a real product link from. Genuine matches for the kind of product you asked for stay ahead of accessories and other off-type results in the base ranking. Each group is ranked on how many sold recently, the star rating, the number of reviews and the price. Your saved Preferences set how much each of those four counts, so the same pool can put a different product first for someone who chose Lowest price or Premium quality; leave them alone and you get the same ranking everyone else gets. When you give answers, it checks titles against your request and answers across the whole ranked pool before keeping ten products, with matches ahead of unclear listings and contradictions removed. A star rating counts in proportion to how many reviews back it, so a perfect score on a handful of reviews carries little weight; a listing under 100 reviews is marked unproven and disfavoured from winning on stars alone, and within its group it cannot head the base ranking while a proven listing there carries ten times its reviews. It can still appear and can still be selected when the model judges it the best fit. It reads customer reviews for the finalists, or Amazon’s own review summary when the fallback returns one, then may name any product in the shortlist as the winner. If the winner came from further down the list, it fetches that listing before showing the card. When none meets everything you asked for, it names the closest and says above it what it could not confirm.
Why did it ask me questions before showing results?
Because your answers shape both the search and the final shortlist. It asks one question at a time about the choices that split this kind of product, such as size, material or type, and writes each next question for the path through the questions so far. You can skip one; the next question then follows that question’s recommended branch, while your search leaves the skipped property open. While you choose, it may prepare the Amazon search made from the first three recommended answers. When you continue, it uses your actual answers; if they make the same phrase, it can reuse the prepared ranking instead of repeating that lookup. It asks three questions before offering the search and can offer up to six; a request that already names the exact product gets none.
Why does it say the winner is not an exact match?
It always names one product. When nothing Amazon returned meets everything you asked for, it picks the closest one and says so in a line above the winner: This is not an exact match. That line continues with one sentence saying what it could not confirm from the listings, and then a link to search Amazon yourself for the same phrase. If you named a medical condition or a specific use and no listing claims it, that sentence says so; read it before you buy. You get no product when the search cannot produce a usable screened shortlist.
Can it be wrong?
Yes. An AI model writes the verdict, and it can misread a listing or miss a better product that the search never surfaced. It is a starting point, not a verdict you should stop thinking at.
Is the price it shows the price I will pay?
Not necessarily. The finalists’ price, rating and exact review count are re-fetched from Amazon when we write the answer, and so are the winner’s when the pick came from further down the list. The prices and purchase-volume figures in More Options come from the initial Amazon search and can be up to a day old. Amazon still changes prices constantly, and the price on Amazon when you check out is the one you pay.
Do you make money when I buy something?
Yes. As an Amazon Associate we earn from qualifying purchases, which means Amazon pays us a small commission on most of what you buy through one of our links. Something you return or cancel earns us nothing. It never costs you more, and it never changes which product wins. The affiliate disclosure has the detail. TwelveTabs is not affiliated with, endorsed by, or sponsored by Amazon.com, Inc.
Would a commission change what it recommends?
No. Amazon pays us a small commission when you buy through our links, and that is how TwelveTabs earns. Sponsored listings are dropped before anything is ranked, nobody can pay for a better ranking, and the commission never changes which product wins. See the affiliate disclosure for how links and ranking are kept separate.
Does TwelveTabs log into my Amazon account or buy things for me?
No. TwelveTabs never asks for your Amazon password, never logs into your Amazon account and never buys anything for you. It finds the product, then sends you to Amazon, where you check out yourself.
Do you keep what I search for?
Yes. The search you entered, the answers you give to our follow-up questions, and both the research and the recommendation or refusal we produced are stored so we can see where the app confuses people. The privacy page lists everything that is kept and who else sees it. Email info@twelvetabs.com and we will delete yours.
Why was I asked to prove I am a person?
Once, on the first search we see from you, and not again while your browser keeps the cookie that tells us it is you. Clear your cookies, or search in a private window, and we cannot tell you apart from someone new, so you will see it again. Every search costs real money, and that check is the only thing stopping a script from opening an unlimited number of fresh identities and spending the day’s budget. You will also see it once more after we delete your data, because that removes the record saying we have seen your cookie before.
It said it was at capacity for today. What does that mean?
One of three daily limits was reached, and the screen tells you which: your own searches for the day, the searches for everyone on your network, or the whole tool’s research budget. Every search costs real money, so it stops rather than spending past a limit. All three reset at midnight UTC, which is the early evening on the US East Coast.
Does it search anything other than amazon.com?
No. It searches the United States Amazon store only, and every link it gives you goes there. Other retailers and other countries are not supported.
Is there an app or a browser extension?
Not yet. The website is the whole product today, and it works on a phone.
Who is behind it?
TwelveTabs is an independent project, not a company, and it is not affiliated with Amazon. This is an early version shared with friends and family, which is why it may change under you. Reach us at info@twelvetabs.com.