Type a keyword - or paste a Home Depot category or product URL - and the listings come back as a table: the product, its brand, its price and list price, its rating and review count, its model and SKU, a link, an image and its availability.
One required field and one setting. There is no sort control and no store or region selector.
The limit is per query, so a list of keywords multiplies it. A broad category with no limit can run long.
What a Home Depot listing row gives you that a screenshot of a search page does not.
price and original_price are separate columns, so a discount is a subtraction rather than a guess. Track the pair over time and you have the shape of a promotion - when it started, how deep it went, and whether it came back.
model and sku both come back. Two identifiers give you two chances to match a row against your own catalogue or against another retailer's feed, which is usually the difference between a join that works and one that quietly drops half its rows.
You are not forced to choose an input shape. A search term, a category page and a single product page can sit on consecutive lines of the same run, which means one job can cover a broad sweep and a short watch-list at once.
One is a hard requirement, one is about how much this page is willing to promise, and one is a quiet trap in the file uploader.
This service needs your own paid residential proxy. When a run comes back with nothing, the platform says why: homedepot.com does not let free proxy addresses through, a paid residential address configured as PROXY_URL is required, and your real IP is never used. That message only appears after an empty run, which is exactly why it is worth reading here first - otherwise an empty table looks like a broken service rather than a missing setting.
We have not measured this export, and this page does not pretend otherwise. The twelve columns below are the ones the service declares, in the order it declares them. What is in them - whether price is a number or a string, what vocabulary availability uses, whether rating and reviews are plain figures, how often original_price is filled - is not stated anywhere on this page, because we hold no export of this service to check it against. Other pages here do state formats; they had runs behind them. Take a small first run and read your own header row before you build a pipeline on any assumption about shape.
The file uploader reads less than you might expect. It offers CSV, XLSX, TXT and Parquet, but a Parquet file is refused in the browser with a note telling you to paste the URLs or use one of the other three. For a CSV or TXT it takes only the first column of each line, so a spreadsheet with the URLs in column B will upload as a list of whatever is in column A. An XLSX behaves the same way against the first sheet. Anything it does read is merged with what is already typed in the box and de-duplicated, which is convenient once you know it and surprising when you do not.
This is the service's own column list, in its own order. The descriptions say what each column is for; they deliberately do not say what its values look like.
url as the destination.price to get a discount. We cannot tell you how often it is filled, because we have not run this service.sku, so the two are not the same number, but which numbering each one uses is not something this page will assert without an export in front of it.model. Useful as a fallback join key; see the note on model for why neither is described more precisely here.These are declared columns, not measured ones. Every other claim on this page is drawn either from the service's own column list or from the platform interface you can see for yourself; none of it comes from an export, because we do not hold one for this service. That is why nothing above states a type, a format, a vocabulary or a fill rate - a plausible guess about availability or price would be worth less to you than an honest gap. Two things you can rely on regardless. The column list itself is fixed at twelve and is what a JSON or Excel download is built from, so a header row will not surprise you. And the run will return nothing at all unless a paid residential proxy is configured first, which is the single most common reason this service looks empty. Take a small first run, read your own header row, and treat that as the specification for anything downstream.
All of them start with something you would type into Home Depot's own search box.
Paste a category URL, run it on a schedule, and keep price next to original_price. The gap between the two is the promotion, and watching it across a whole category tells you more than watching any single product does.
Upload your keywords or product URLs, then join the results back to your own catalogue on model or sku. Two identifier columns means a row that fails to match on one still has a second chance on the other.
A broad category run with no limit returns the assortment as rows - brands, price points, ratings and review counts together. That is a different question from what any one product sells for, and it is the one worth asking first.
The same flat rate as every other scraper on the platform. No plan, no seat, no monthly minimum.
One-time on every new account. No credit card. Every scraper unlocked - and enough to confirm your proxy is configured and to read this service's real header row for yourself before you build anything on it.
About $2 per 1,000 products, the same flat rate as any other scraper here. The limit applies per query, so a list of keywords multiplies it - and leaving it empty takes everything.
Volume pricing, an SLA, dedicated workers and bespoke integration for teams tracking retail assortments on a schedule rather than checking a handful of products. Tell us your numbers and we will quote.
LIVESCRAPER10 at checkout.This one returns the product row. These return what is behind it, or the same shelf at a competitor. Each has its own schema.
The natural next step: this page gives you products and their review counts, that one gives you the reviews themselves. It is a separate service on a separate endpoint with its own column list - read both dictionaries rather than assuming they line up.
The same shelf at the nearest competitor. Running a category on both is how a price comparison stops being anecdotal - though the two services have their own schemas, so match on your own identifiers rather than on column position.
A third home-improvement retailer, useful when a two-way comparison keeps coming out ambiguous. Same caution as above about schemas: treat each service's column list as its own.
Short answer: these are public product listings, and nothing in the twelve columns describes a person.
Everything in these columns is shown to any visitor on a public homedepot.com search or product page. No login is used, no paywall is crossed and no account is touched. Collecting publicly listed product and price information for research and comparison is long-established practice, and price transparency is one of the clearer cases for it.
This export carries no personal data. Every column describes a product - a name, a brand, two prices, a rating and a count, two identifiers, a link, an image and an availability state. There is no reviewer, no seller name, no contact and no email anywhere in the declared list, which is not something we can say about most pages on this site. The rating is an aggregate of what customers said rather than the things they said; if you want the individual reviews, that is the separate service linked above, and it carries a heavier obligation.
The requirement to route through your own paid residential proxy is a technical condition of use rather than a legal one, but it is worth reading as a signal: Home Depot puts real effort into controlling automated access, and its terms govern it. Read them before you scale, keep collection proportionate to the research you are actually doing, and do not treat published prices as a licence to republish a catalogue wholesale. We run no third-party trackers on the data layer, and your exports auto-delete after 30 days.
The two return very different things, and running both is the quickest way to see what this export is good for. Your first 500 rows are free.
The Home Depot Products Scraper turns a homedepot.com search into a spreadsheet. You submit keywords or Home Depot URLs - one per line, mixed freely, category pages and single product pages both accepted, or uploaded as a CSV, XLSX or TXT file - set a limit per query, and each listing comes back as a row: the product name, its brand, its price and its list price, its rating and review count, a model number and a SKU, a link to the page, an image and an availability state. Twelve columns, one row per product, downloadable as JSON or Excel.
Two things shape how you should use this, and both are worth knowing before a first run. The first is that the service requires your own paid residential proxy, configured as PROXY_URL. Home Depot does not let free proxy addresses through, and a run without one comes back empty - the platform explains why only after the fact, which is why it is stated up front here. The second is that this page documents the declared column list and deliberately stops there. We hold no run export for this service, so nothing here claims a data type, a value format, a vocabulary for availability, or how often original_price is populated. Several other pages on this site do state measured formats, because runs were available to read; this one is honest about not having them. Take a small first run and let your own header row be the specification.
What the column list does tell you is worth having. price and original_price arrive as separate columns, so a discount is arithmetic rather than inference, and tracking the pair across a category over time gives you the shape of a promotion rather than a snapshot of one. model and sku are declared separately, which means two chances to join a row against your own catalogue or a competitor's feed - usually the difference between a match rate you can trust and one that silently drops rows. And because keywords and URLs share the same input box, a single job can cover a broad category sweep and a short watch-list of specific products at once.
A few practical notes. The limit defaults to 100, accepts a minimum of 1, and applies per query rather than per run, so a list of keywords multiplies it; leave it empty to take everything, though a broad category with no limit can run long. The file uploader reads only the first column of a CSV, TXT or XLSX, and refuses Parquet in the browser, so keep your URLs in column A. There is no sort control and no store or region selector - sort the export after you download it. On the legal side this export is unusually clean: every column describes a product rather than a person, and no reviewer, seller, contact or email appears anywhere in the declared list. Home Depot's terms still govern automated access, so read them before you scale. Start free: your first 500 rows cost nothing and need no credit card, then $0.002 per row, flat. See pricing for current rates.