How this works
A shopping site that is willing to
tell you not to buy.
Gadgetly is built around one idea: you should not have to translate what you need into product names, filters and price columns. Describe the job, and the platform does the translating — visibly, so you can argue with it.
The question is what you should do
A marketplace answers where to buy. A comparison site answers what is cheaper. A review answers what one thing is like. None of those is the question you actually arrived with.
Every number shows its working
A match percentage always comes with the dimensions, the weights you set, and the product score in each. If we cannot explain a ranking, we do not show it.
Commission is never an input
Ranking runs on scores, price behaviour and your stated priorities. No retailer can pay to move up a list, and the code that ranks has no access to commercial terms.
We say when we do not know
Conditional discounts are labelled conditional. Modelled figures are labelled modelled. Prototype data is labelled prototype, on every screen it appears on.
The method
Five steps between “I need something” and a decision you trust.
Most shopping sites do one of these and leave you to assemble the rest. The value is in the chain, not the links.
01
Intent
You describe the job in your own words. No category tree, no filters to learn.
02
Requirements
We turn that into a budget, a use case and a set of weighted priorities — and show you every one, editable.
03
Match
Products are scored against your weights, not against a generic average. The reasoning is printed on the card.
04
Real price
Every store, every card offer and coupon, resolved into the number you will actually pay.
05
Timing
Against its own price history: buy now, wait, watch, or skip entirely.
Ranking is driven by scores, price and your stated priorities. Retailer commission is never an input.
Try it yourself
Every claim on this site is a control you can move.
These are not screenshots. Each panel runs the same code the product runs on, in your browser, right now.
Type anything. Watch it get read.
No dropdowns, no filters. This is the real parser running in your browser — budget, category, use case and priority weights, extracted live. Every one of them is editable afterwards.
Category
Laptops
Budget
Under ₹90K
Use
Gaming + College
Priority
Performance, Value
Priority weights it derived
Compared with
What everywhere else answers instead
Every alternative answers a different question from the one you arrived with.
A marketplace
Amazon, Flipkart
“Where can I buy it?”
Assumes you already know what you want. Sorts by what it can sell you.
A comparison site
Price aggregators
“Which one is cheaper?”
Compares sticker prices, not what you will actually pay after card offers and coupons.
A review site
Tech blogs, YouTube
“What is this one like?”
One product at a time, reviewed for everyone. Never priced, never timed.
A chatbot
General AI
“Anything, confidently.”
No live prices, no store list, no price history. It cannot tell you to wait three weeks.
Every one of those answers a different question from the one you arrived with, which is what should I do?
So this is what Gadgetly does
You describe the job, not the product
“Gaming laptop under ₹90K for college” is a sentence, not a filter tree. We read the budget, the use case and what you are trading off — and show you every assumption we made.
Every recommendation shows its working
A match score always arrives with the dimensions, your weights and the product score in each. Move a slider and the shortlist reorders in front of you. Nothing is hidden behind an algorithm you cannot see.
The real price, not the sticker price
Every store, every card offer and coupon, resolved into one number — with the conditional ones flagged so you can switch off the discount you cannot claim.
It will tell you not to buy today
Priced against its own history, we say buy, wait, watch or look elsewhere. A site that never says wait is selling, not advising.
It builds the whole setup, not one box
Most people are furnishing a desk, not buying a product. Give us the goal and the ceiling and we allocate the budget across every slot, then show which single upgrade is worth the leftover.
Commission never moves the ranking
We earn when you buy through a retailer link. The code that ranks has no access to commercial terms, and the disclosure sits next to the button rather than in a footer.
Scoring
Ten dimensions, one scale
Every product is scored 0-100 on the dimensions that apply to it. The same numbers drive the plain-language spec view, the match score and the comparison table — so those three can never contradict each other.
In this prototype the scores are illustrative editorial values written to demonstrate the product experience. They are not the output of independent laboratory testing, and the interface never claims otherwise.
Deal score
Why we ignore the discount percentage
The MRP a discount is measured against is set by the seller. That makes “45% off” a marketing decision rather than a measurement, and it is the single most misleading number in online retail.
Our deal score is weighted 45% on where today's price sits inside the product's own observed range, 30% on the median price over that window, 15% on what other stores are asking right now, and only 10% on the discount off MRP — capped, so a padded MRP cannot carry the score.
See it appliedEffective price
The sticker price is rarely the price
Card offers, coupons, exchange values and cashback move the real number substantially, and they do not all apply to everyone. Every offer breaks the deductions out separately and marks which depend on your eligibility.
Exchange and cashback are excluded from the headline figure by default. An exchange quote changes after inspection, and cashback that arrives in 60 days locked to one store is not the same as money off today.
Read the guideAccuracy
What we promise, and what we cannot
Retail prices move hourly. Bank offers are set by banks, capped, frequently limited to specific cards, and can change or end without notice. No comparison site on earth can guarantee that a number it shows will still be true when you reach checkout, and any site that implies otherwise is misleading you.
So we do not promise accuracy. We promise transparency, recency and correction speed. That means four concrete things:
Every number is timestamped
Prices and bank offers are checked on different schedules and carry separate timestamps. Anything stale is labelled stale rather than shown as if it were fresh.
Bank offers show their expiry
Each one displays when we last confirmed it, where it came from, and the date it runs out. An offer without an end date is not a fact we have.
Conditional means conditional
A capped card discount is never folded silently into one headline price. It is a separate line, marked "if eligible", and you can switch it off.
Anyone can report a wrong number
Every offer carries a report control. A report flags the entry for re-checking, and the retailer's checkout page always overrides ours.
The price you pay is whatever the retailer shows at checkout. Treat everything here as a shortlist that saves you opening nine tabs — not as a quote.
Data
Where the numbers in this build come from
This is a frontend prototype. Being precise about that is part of the point.
A seed catalogue of 38 products across ten categories, written for this build. Names and specifications are realistic; the scores are editorial illustrations.
Generated deterministically for each product across six retailers, including card offers, coupons and stock states. No live retailer feed is connected.
Synthetic series, up to 180 days per product and never longer than the product has existed. Shaped to match each product’s editorial verdict.
What changes when this goes live
The interface reads everything through a data layer — catalogue, offers, pricing, matching, setup building — rather than touching the seed files directly. Replacing the seed data with a CMS and live retailer feeds means changing those modules and nothing else. No screen in this build would need rewriting.
Affiliate disclosure
How the money will work
In production, some links to retailers will earn a commission when you buy. That is how a site like this pays for itself without charging you or selling your data.
Two commitments come with that. First, commission is not an input to any ranking, score or verdict — the ranking code has no access to commercial terms. Second, the disclosure stays visible next to the outbound button, not buried in a footer.
In this prototype, no affiliate links are active at all. Every “Go to store” button opens an explanation instead of a redirect, and no tracking parameters are attached to anything.