momox second-hand

Search or browse either storefront with momox's own search engine. One row per item: price, condition grade, brand and — on fashion — size, exact colour and colour family, material, length and the garment's own feature list, plus its shelf parsed into gender / category / subcategory. Free-text `q` and `category` can be combined, which the group's media storefronts refuse to do, and every filter is applied BY THE SOURCE, so the counts that come back are momox's own: `jacke` 170,726 items, narrowed to 13,135 inside men's jackets, to 475 for Nike and to 151 for Nike in size INT M. Two totals are always returned because the source publishes two: `total_matched`, the real match count, and `total_window`, the 10,000 rows actually reachable by paging. Filter values are case-sensitive at the source and a mis-cased one comes back as 0 rows with HTTP 200, so each value is folded to the storefront's own spelling and then checked against the source's own echo of what it applied; anything it did not apply is reported rather than passed off as a narrowing.

Everything one item publishes. The complete offer table straight from the shop's own product endpoint: on books one offer per condition grade with its own price and its own stock (ISBN 9782070584925 publishes four — like new 7.49 sold out, very good 5.79, good 4.01, acceptable 3.99 — and all 12 titles measured published more than one), on fashion the single offer behind the single garment, which is all there is: 12 of 12 garments measured carried exactly one. `price_min`, `price_min_available` and a per-grade price table are all returned, because the cheapest grade is sometimes the sold-out one. On fashion the item's page is read as well for the things only it carries — momox's OWN measurements of that garment in centimetres, its exact colour, its material list, its feature list, its condition grade in momox's words and its article number. The page's schema.org block is read as an independent price witness on both storefronts and any disagreement is reported in `meta.schema_org_cross_check` rather than hidden: it agreed with the live price on 12 of 12 garments, and on books it quotes the storefront's DEFAULT grade, which was the cheapest published offer on only 3 of 12 titles — a third number with a third meaning, explained instead of flattened. Reachable by handle, by a full product URL, or on books by ISBN-13, ISBN-10 or EAN. An unknown handle is NOT_FOUND.

momox's own inventory census: every value its search engine offers for a query or a shelf, with ITS OWN item count beside each one. This is how you learn the vocabulary before filtering, and it is the only honest source of it, because the source matches filter values case-sensitively and answers a mis-spelled one with 0 items and HTTP 200. Fashion publishes 14 facets plus price — 187 size values for the keyword `jacke`, 1,848 brands, colour family, condition, material, length, fit, neckline, sleeve length, pattern, closure, fabric texture, details, heel height — and books 7: condition, author/artist, format, platform, language, medium and in-stock. It is also a census in its own right: `jacke` reports like new 68,528 against very good 100,895 and good 27, which is momox's grading mix for jackets in one number.

momox's whole category tree with its own item count on every node and the Shopify collection handle that serves it. On fashion the node id IS the structured token the shop stamps on every product, so the tree, the search filter and each row's own `category` field are one and the same vocabulary: All 1,854,875 -> Damen 1,515,748 -> Bekleidung 1,377,790 -> and down to Handtaschen 8,724. On books it is All 12,130,377 -> Books 10,460,867 -> Professional & Reference 4,385,001 and so on. Walk from any node with `root` and control the depth; depth 1 is the cheapest way to see the shape of the catalogue. These counts are the source's own index counts, not a live buyable count — the `collections` action's numbers are a different and less reliable thing, and both are labelled as what they are.

Every collection the Shopify storefront publishes, with handle, title, description, image and ready-made `feed_params`. This is the list of shelves the shop's own navigation uses, and the handles are what `latest` reads a feed from. The count the source attaches to each one is published as `products_count_claimed` and nothing else, because it is not a live number: momoxbooks publishes hundreds of `by-<artist>` collections whose count is 0, and the Shopify count includes products that can no longer be bought. Use `categories` for a count you can rely on, or `search` for the live match count.

The shop's live product feed, newest first, straight from its own catalogue endpoint — 250 items per request with the authoritative price, the crossed-out price where there is a real one, the stock flag and every image. This is the surface that settles the price question: 2,000 of 2,000 fashion rows measured carried a real price, while 1,992 of them carried `compare_at_price` as the string "0.00" (dropped to null here), and on books the 105 of 1,017 variants whose price is "0.00" are exactly the ones with no copy in stock, published as `offer_status: no_copy_in_stock` and a null price rather than as a free book. Use it to watch what momox listed today — page 1 was published the same morning it was measured — or to read one Shopify collection's feed. Note that this endpoint honours only `limit` and `page`: every filter and sort parameter it advertises was measured returning the identical 50 product ids in the identical order, so all real searching belongs in `search`.

The storefront's own autocomplete: what momox suggests as you type. Returns both the suggested SEARCH TERMS (`nike` -> `nike damen`), each with ready-made `search_params`, and the matching ITEMS in full — on fashion already carrying price, condition, exact colour, material, size and the taxonomy token, so a single cheap request answers 'what does momox have for this phrase' without a full search. One upstream request, no wall in front of it.