Live-commerce and category trend tracking
Whatnotエンドポイントを使えば、「Live-commerce and category trend tracking」をドキュメント化された入力とJSONレスポンスによる再現可能なAPIリクエストに変えられます。
Turn Whatnot's public live-shopping data into structured JSON — browse live shows by category, the full category list, and a specific live show's current shop feed (items up for sale, pricing), as normalized JSON. Credential-free.
Browse Whatnot live shopping shows by category and get a live show's current shop feed as structured JSON.
エンドポイントファミリー
4
ドキュメント化されたパラメータ
8
サンプル数
4
ライブカタログスナップショット
稼働エンドポイント
4
メソッド
GET
必須パラメータ
7
スキーマ参照
4
{
"platform": "Whatnot",
"endpoint": "whatnot-browse",
"method": "GET",
"path": "/whatnot/browse",
"auth": "apiKey"
}ユースケース
Browse Whatnot live shopping shows by category and get a live show's current shop feed as structured JSON.
Whatnotエンドポイントを使えば、「Live-commerce and category trend tracking」をドキュメント化された入力とJSONレスポンスによる再現可能なAPIリクエストに変えられます。
Whatnotエンドポイントを使えば、「Live show and seller monitoring」をドキュメント化された入力とJSONレスポンスによる再現可能なAPIリクエストに変えられます。
Whatnotエンドポイントを使えば、「Auction pricing research」をドキュメント化された入力とJSONレスポンスによる再現可能なAPIリクエストに変えられます。
マネージド実行
以下の数値はすべて、稼働中のWhatnotエンドポイントカタログから取得しています（エンドポイント4件、ドキュメント化されたリクエストパラメータ8個、公開レスポンススキーマ4件）。DocsとPlaygroundが参照しているカタログと同じものです。
Whatnotのエンドポイントは4件、4つのリクエストファミリーに分類されています（Browse、Categories、Liveほか1件）。
これらのWhatnotエンドポイントには8個のリクエストパラメータがドキュメント化されており、うち7個が必須です。統合コードを書く前に入力仕様をすべて確認できます。
4件のWhatnotエンドポイントのうち4件が実際のサンプルレスポンスを、4件がドキュメント化されたレスポンススキーマを備えています。最初のリクエストの前に実データの形に合わせて実装できます。
Whatnotのエンドポイントは成功時のスキーマに加えてエラーレスポンス（400、404、500、503）もドキュメント化しています。ブロック、レート制限、レコード欠損は空データではなく型付きのエラーとして返ります。
Whatnotのエンドポイントは4個のホスト型MCPツールとして提供されており、エージェントは同じパラメータと同じJSON契約のまま、追加の実装なしに同じルートを呼び出せます。
カバレッジマップ
/whatnot/browse
/whatnot/categories
/whatnot/live/{id}
/whatnot/seller/{username}
エンドポイントカタログ
/whatnot/browseReturns the live and upcoming shows currently listed under a Whatnot category: seller, title, status, start time, thumbnail, and tags.
レスポンスに関する注記
- `id` is a show's live-stream id — pass it directly as the `id` path parameter to [`/whatnot/live/{id}`](whatnot-live.md) for that show's current shop feed. - `status` reflects Whatnot's own show state: `PLAYING` (currently live) or `CREATED` (scheduled, not yet started — a show that hasn't started yet still appears in the list with a future `start_time_ms` and a `CREATED` status). An ended-show state almost certainly also exists but was not observed in live testing — ended shows drop out of Whatnot's own Browse/ category feeds essentially immediately. - An unrecognized `category` returns `404`. - A missing `category` returns `400` before any upstream request is made.
MCPツール whatnot_browse
/whatnot/live/{id}Returns a Whatnot live show's current shop feed: every product, auction, and giveaway listing currently visible in the show, each with its seller's rating.
レスポンスに関する注記
- `transaction_type` distinguishes how a listing is sold: `AUCTION` (bid-based, `current_bid_cents`/`current_bid_count` populated once bidding starts), `GIVEAWAY`, or `BUY_IT_NOW`. - `status` reflects the listing's own lifecycle, not the show's overall live/ended state: `active`, `created`, or `running`. A terminal "sold"/"sold out"-style value almost certainly also exists but was not observed in live testing. - `seller.rating`/`seller.review_count` are scoped to the listing's seller, not the requester — there is no separate seller-profile lookup this endpoint calls. - An unknown or ended show id returns `200` with `"products": []`, not an error — Whatnot's own backend doesn't distinguish "no such show" from "a real show with nothing currently listed." - A missing `id` returns `400` before any upstream request is made.
MCPツール whatnot_live
/whatnot/seller/{username}Returns public seller profile details and one page of the seller's livestreams. Use next_cursor as cursor to continue while has_more is true. The endpoint does not include shop products.
MCPツール whatnot_seller
/whatnot/categoriesReturns Whatnot's full top-level category list (e.g. "Trading Card Games", "Sneakers & Streetwear"). Each entry's slug is usable directly with /whatnot/browse's category filter.
レスポンスに関する注記
- Each entry's `slug` is the value to pass as [`/whatnot/browse`](whatnot-browse.md)'s `category` query parameter. - This list changes rarely; results are fetched live on every call, not cached client-side.
MCPツール whatnot_categories
関連API
Marketplaces & Retail
Research new, used, and vintage instruments across Reverb listings, seller shops, and the Price Guide. Keep current asking prices separate from historical sales, and preserve model, finish, year, condition, and currency when comparing gear.
Marketplaces & Retail
Collect marketplace product signals from Amazon without building brittle storefront scrapers.
Marketplaces & Retail
Build resale, pricing, and marketplace workflows from structured eBay data.
Whatnotのスクレイピング方法
Crawlora's Whatnot endpoints return category browsing, a live show's current shop feed, and public seller profile data with public livestream cards as normalized JSON using a Crawlora API key. The new /whatnot/seller/{username} route is a profile-and-show archive, not a shop product listing or seller search: it returns at most 15 shows per call and uses an opaque cursor copied unchanged from next_cursor while has_more is true. Preserve total_count as a changing source count, and do not assume that a profile's show list includes every ended stream.
/whatnot/categories takes no parameters and returns all 37 categories, each with an id, a display name and a slug. You need the slug, not the name — it is underscore-separated, so Sports Cards is sports_cards.
Pass that slug to /whatnot/browse — the display name or a hyphenated variant returns a 404, so sports_cards works while "Sports Cards" and sports-cards both fail. A browse returns around 24 shows, each with id, url, title, seller_username, status, start_time_ms and thumbnail.
status tells you whether a show is live now or scheduled, and start_time_ms is a Unix timestamp in milliseconds — divide by 1,000 before feeding it to a seconds-based date function.
Pass a show id from /whatnot/browse to /whatnot/live/{id} for that show's current shop feed — items up for sale and pricing.
Pass a Whatnot username to /whatnot/seller/{username}. The response combines public profile fields with at most 15 public livestream cards; follow next_cursor unchanged only while has_more is true. It does not return product listings or add sorting/filter parameters.
Re-run browse or a specific show on a schedule to track category activity and item pricing over time. Because a show's feed changes while it runs, sampling interval matters more here than on a static catalog.
On 2026-10-30 /whatnot/categories returned 35 categories, browse for trading_card_games returned 24 shows with status and start_time_ms, and the seller fluffcardhouse returned 14 shows with has_more false.
エージェントスキル
Crawlora のエージェントスキルをインストールすると、スクレイパーを書かずに、ドキュメント化された API 経由で Whatnot のデータを JSON として取得できます。Whatnot に対応するスキル:
Researches secondhand, resale, handmade, and musical-instrument marketplaces via the Crawlora API. Use to compare live listings and sellers, check…
npx skills add github.com/Crawlora-org/crawlora-skills --skill resale-secondhand-researchFAQ
Fetch /whatnot/categories first, then pass a category's slug to /whatnot/browse. The response is a normalized list of current live shows with seller, status, start time and thumbnail — no Whatnot account required.
Almost certainly because you passed the category's display name rather than its slug. The slug is underscore-separated: sports_cards succeeds, while both "Sports Cards" and sports-cards return a 404. Every valid slug is listed in /whatnot/categories, so read them from there rather than deriving them from the names.
Yes — /whatnot/live/{id} returns a live show's current shop feed, including items and pricing, as structured JSON.
No. /whatnot/seller/{username} returns public profile fields and one page of public livestreams, up to 15 per call. It has no product-search, filtering, or sorting controls; cursor paging uses the exact next_cursor from the previous response.
No Whatnot login or auth token is required from the caller — only your Crawlora API key.