Products

amz product is the workhorse. It fetches a detail page and normalizes it into one record that carries everything the page exposes.

There is no structured blob to read first. amazon.com publishes no JSON-LD, no __NEXT_DATA__ and no Apollo cache, so every field comes from a region Amazon named, a JavaScript payload it ships, an attribute, or a selector, and the record says which of those answered.

One product

amz product B075F5X8BR
amz product B075F5X8BR -o json

You can pass several at once, or full URLs; amz extracts the ASIN from any Amazon URL shape:

amz product B075F5X8BR B07XJ8C8F5 "https://www.amazon.com/dp/B08N5WRWNW"

The fields

A product record names every field the page had:

Field	Meaning
asin, parent_asin	the item, and its variation parent when present
title, byline, manufacturer	identity as the page words it
brand, authors	references to the storefronts behind those names
isbn10, isbn13, upc, ean, model_number	the identifiers the detail table carries
offer	the buy box: price, list price, savings, coupon, availability, seller, delivery, returns
other_offers	how many other buying options the page claims
rating, ratings_count, reviews_count	social proof
distribution	the five bucket star histogram
reviews, questions	how many exist, not what amz holds: loaded, total_count, complete
review_sample, qa_sample	the reviews and pairs amz does hold, which is what the detail page embedded
bought_past_month, bought_past_month_n	the "N+ bought in past month" line, and the parse of it
description, bullets	marketing copy
details	the technical details table, as key and value pairs
images, image_urls, videos	one entry per distinct photo with its variants, the same photos as plain URLs, and the inline videos
breadcrumb, ranks	browse node references, and every Best Sellers Rank line
variation	the twister: dimensions, the current selection, and the siblings
rails, similar_asins	the recommendation strips and the similar items
extra	every payload the page shipped that no field reads yet, verbatim
envelope	which responses this came from, what answered each field, and what was looked for and not found

Fields the page did not carry are omitted, and that omission is load bearing: if a field is missing and nothing in envelope.missed names it, amz read the place that field lives and there was nothing there.

Prices are objects rather than numbers. Each keeps the string Amazon printed beside the parse of it, so a wrong parse is recoverable, and a product with no price serializes as null rather than as 0.

The buy box

offer holds price, list_price, savings, savings_pct, per_unit and per_unit_label, coupon, subscribe, business_price, condition, availability, in_stock, sold_by, ships_from, prime, delivery, and returns.

savings and savings_pct are computed from the two prices rather than read off the page. Amazon prints its own saving line and the two disagree often enough that the derived pair is the honest one. availability is the line as written and in_stock is amz's reading of it, and both are kept because the reading is a guess about phrasing and the string is not.

Images and videos

Amazon serves the same photo at dozens of sizes and from several CDN hosts. amz strips the size modifier from every image URL and pins one canonical host, so images holds one full-resolution URL per distinct photo, with the thumbnails, tracking pixels, and sprites removed. The same canonicalization runs on every surface that carries an image (search cards, reviews, brand logos, author photos), so an image URL means the same thing everywhere.

Ranks

A product is usually ranked once overall and again in one or more subcategories. Each entry in ranks carries rank, category, node, and overall:

amz product B075F5X8BR -o json --fields asin,ranks

overall marks the department level rank, the one people mean by "sales rank", and it is flagged rather than assumed to be the first line. node is the browse node the rank line links, which is the strongest category edge amz has, because it is an identifier Amazon stated rather than a name lifted from a breadcrumb.

Variations

variation carries the twister: the parent ASIN, the dimensions Amazon offers, the current selection, and the siblings. Add --variants to print the sibling ASINs as their own rows:

amz product B075F5X8BR --variants -o jsonl

A sibling always has its ASIN and its dimension values. It gets a price, an image and an availability only at --depth deep, which fetches each sibling's own page, because the parent page states none of those three and filling them from the parent would quote the selected variant's price against every colour in the family.

Depth

--depth decides how much is read:

Depth	Requests	What you get
quick	1	the mobile URL, /gp/aw/d/, which returns the same page as meta
meta	1	the full 2.2 MB detail page, rails dropped
full	1	the same page with the recommendation rails kept
deep	2 plus one per sibling	full, plus each variation sibling's own page and the seller

meta is the default. It drops the rails it read and records that it did, so a product with twelve strips on the page never looks like a product with none. full costs no extra request, it just keeps them.

quick was specified as the cheap read and is not one. Amazon gzips both responses, so the 374 KB it was specified at is what the mobile URL weighs on the wire while the 2.2 MB it was compared against is what the detail page weighs after decoding. Measured on 2026-08-17 for B075F5X8BR, /gp/aw/d/ is 373,945 bytes on the wire and 2,197,291 decoded, against 373,980 and 2,196,553 for /dp/. It is the same page for the same bytes. quick is kept because it is the only thing that reads that surface, and the saving returns if Amazon serves it a lighter rendering again.

deep prints its bill and stops for --yes above twenty requests. An apparel listing with 88 siblings is ninety requests and a minute and a half of polite crawling for one record, and that is worth agreeing to rather than discovering.

amz product B075F5X8BR --depth quick     # cheapest useful read
amz product B075F5X8BR --depth deep --yes

The buy box alongside

--with-offers emits the buy box as its own offer row beside the product, so a pipeline that wants both shapes gets them from one fetch:

amz product B075F5X8BR --with-offers -o json

It is the buy box and not a list of every seller. The other sellers' rows are drawn by JavaScript and the endpoint behind them answers 404 to a direct request, so what the page states is a count, and that count is in other_offers. See offers for the measurement.

The raw page

When you want the bytes amz parsed, not the record:

amz product B075F5X8BR --raw > page.html

Just the price

For price-watching, price skips everything else:

amz price B075F5X8BR B07XJ8C8F5
amz price B075F5X8BR -m uk

Recommendation rails

related pulls the recommendation cards off a detail page, the "customers also viewed" and "frequently bought together" rails:

amz related B075F5X8BR
amz related B075F5X8BR --kind also-viewed -o jsonl

Dry run

See the URL without fetching, useful when scripting across marketplaces:

amz product B075F5X8BR -m de --dry-run