How HAULS tracks prices and discounts

Where the data comes from

HAULS builds its price history from three kinds of sources, never from one:

• User-observed data, a price seen by the HAULS extension on a real user’s page load, the same way any shopper would see it.
• Licensed or API data, where a marketplace or partner makes it available directly.
• Public data, published structured product information such as JSON-LD/Schema.org markup that a store already puts on its own page.

Amazon’s and Flipkart’s own terms explicitly prohibit systematic scraping or bot-based extraction without permission. HAULS is built around that constraint, not against it, the moat is accurate product identity, price normalization, and transparent reasoning, not bypassing anti-bot systems.

Resolving the actual price

A product page rarely shows one clean number. It shows a selling price, a struck-through MRP, an EMI figure, a refurbished-condition price, and often prices for products the algorithm is merely recommending. HAULS’ price resolver is built to tell these apart, in order of preference:

1. Structured data the page itself publishes (JSON-LD / Schema.org), where available.
2. DOM-level extraction, as a fallback when structured data is missing or incomplete.
3. The product’s URL or ID, to keep price history tied to the right product over time.

Every resolved price is stored as a full event, price, seller, availability, variant, and timestamp, not a bare number.

Confidence, shown honestly

Every price HAULS shows carries a visible confidence level, never a silent guess:

• Verified from product page, confidence: high.
• Price detected from page structure, confidence: medium.
• Price could not be reliably determined, shown as-is, instead of guessing.

HAULS will never display a “lowest price ever” claim it can’t support. If the observation window is short, that’s stated plainly, for example “History available: 12 days,” instead of presenting a short sample as a long trend.

How discount context works

An advertised discount compares today’s price with a reference price the seller chose, usually the printed MRP or a struck-through “was” price. HAULS compares today’s price with what it has actually observed for that product recently instead. When the two disagree, that gap is shown as discount context, a factual comparison, never an accusation that a specific seller inflated a price on purpose.

Matching products across stores

“Nike Air Max 270 Black 9” on one store and “Nike Air Max 270 Men’s Black Size 9” on another are the same product; “Nike Air Max 270 SE” is not. HAULS builds a canonical identity, brand, product family, model, SKU, variant, colour, size, and quantity, and only shows a cross-store comparison once that match is confident. Comparing non-equivalent products would be worse than showing nothing.

What HAULS won’t do

These aren’t cosmetic wording choices, they’re the trust boundary of the product:

• HAULS won’t call a discount “fake”, it shows discount context and lets the numbers speak.
• HAULS won’t claim a review is “fake”, it surfaces unusual review patterns as a signal to notice.
• HAULS won’t accuse a named seller, brand, or marketplace of cheating anyone.
• HAULS won’t sell recommendations, sponsored placement, or affiliate-biased rankings.

Questions