# Target (/docs/target) Target [#target] Target.com product detail, customer reviews, category browsing, and store locations behind the same unified envelope and canonical commerce schema as Amazon, Walmart, and Google Shopping, so one parser handles all four. Base URL: `/v1/target/...` Getting started [#getting-started] Every endpoint is a `GET` with query parameters and an `x-api-key` header. Start from a TCIN, Target's catalogue id, which is the number at the end of any Target product URL: ``` https://www.target.com/p/beats-solo-buds-true-wireless-bluetooth-earbuds/-/A-92148487 ^^^^^^^^ ``` ```bash curl "https://www.socialcrawl.dev/v1/target/product?tcin=92148487" \ -H "x-api-key: $SOCIALCRAWL_API_KEY" ``` Endpoints [#endpoints] Read this before you build [#read-this-before-you-build] **There is no Target keyword search, and that is deliberate.** Target's search surface returns at most 24 products, ignores the page parameter entirely, and answers a query that matches nothing with 24 unrelated products rather than an empty list. There is no way to tell a real result from a miss, so shipping it would mean billing you for rows you did not ask for. **Browse instead.** `GET /v1/target/categories` gives you the whole category tree for 1 credit, and `GET /v1/target/category` walks any node 24 products at a time with real prices and working pagination. That is the reliable path from nothing to a set of TCINs. If you specifically need keyword matching, use [`/v1/search/everywhere`](/docs/universal-search) for a cross-source sweep. **A variant TCIN resolves to its parent product.** Target rolls colours and sizes up: asking for `92146817` (Matte Black) returns the parent product `92148487`, whose payload describes the family. `product.id` always carries the id of the product the payload actually describes, and `product.ext.requested_id` carries the id you asked for whenever the two differ, so the round trip is never lost. Colour and size options come back in `product.variations`, each with its own TCIN. **Target product pricing is national.** Unlike Walmart, Target's product endpoint reports the same price whatever store is asked for, so there is no store or ZIP parameter here. Treat `price.current` as catalogue pricing rather than shelf pricing. Typical flow [#typical-flow] ```bash # 1. Get the category tree (1 credit, cached hard) curl "https://www.socialcrawl.dev/v1/target/categories" \ -H "x-api-key: $SOCIALCRAWL_API_KEY" # 2. Walk a category for TCINs and real prices, 24 per page curl "https://www.socialcrawl.dev/v1/target/category?category_id=hz89j" \ -H "x-api-key: $SOCIALCRAWL_API_KEY" # 3. Full detail on one product, including the 1-5 star breakdown curl "https://www.socialcrawl.dev/v1/target/product?tcin=92148487" \ -H "x-api-key: $SOCIALCRAWL_API_KEY" # 4. Walk its reviews, 10 per page, passing the cursor back until has_more is false curl "https://www.socialcrawl.dev/v1/target/reviews?tcin=92146817" \ -H "x-api-key: $SOCIALCRAWL_API_KEY" ``` Already hold a TCIN? Skip straight to step 3. **De-duplicate a category crawl by `product.id`.** Target's category ordering shifts between calls, so consecutive pages occasionally repeat a handful of products. We measured 0 repeats one morning and 4 of 24 the same afternoon on the identical category. Every row carries a stable id, so a `Set` is all it takes. Reviews do not have this problem: those pages are cleanly separated. Stores [#stores] `GET /v1/target/stores?query=10001` returns the Target stores near a ZIP code, city, or state, each with its store id, composed address, phone, distance, open status, timezone, and two weeks of daily opening hours. It is 1 credit because store data barely changes and caches hard. Target publishes no coordinates for a store, so `latitude` and `longitude` are null rather than guessed; geocode the returned address if you need them. The store-locator extras ride on `place.ext`: ```json "ext": { "distance": 0.37, "status": "Open", "timezone": "America/New_York", "hours": [ { "date": "2026-07-27", "day_name": "Monday", "is_open": true, "opens_at": "08:00:00", "closes_at": "23:00:00" } ] } ``` A store id from here does **not** localise prices: see the pricing note above. Reviews pagination [#reviews-pagination] Target review pages do not overlap and the review total is exact, so a full crawl is deterministic: pass the returned `pagination.next_cursor` back as `cursor` until `has_more` is `false`. Page size is fixed at 10 and cannot be raised. If what you actually want is the rating breakdown rather than the review text, call `/v1/target/product` instead and read `product.ext.rating_distribution`. It carries the same 1 to 5 star counts in a single call, rather than one call per ten reviews. ```json "ext": { "rating_distribution": { "star_1": 549, "star_2": 149, "star_3": 124, "star_4": 148, "star_5": 932 } } ``` Ratings and reviews are two different numbers [#ratings-and-reviews-are-two-different-numbers] `product.rating.count` is the number of people who left a star rating. `reviews_count` is the number who wrote review text, and it is the same population `/v1/target/reviews` paginates over. On a typical product the first number is two to three times the second. Both are reported exactly as Target reports them; neither is an estimate. Response shape [#response-shape] Products and reviews use the shared canonical commerce objects documented under [Schema](/docs/schema). A Target product and a Walmart product have the same leaf names, so a price-comparison integration needs one mapping, not one per retailer. Notes [#notes] * All endpoints use `GET` with query parameters * Authentication via the `x-api-key` header * Responses follow the unified SocialCrawl schema * Empty results and not-found TCINs are refunded automatically and never billed * Prices are US dollars; Target is a US-only retailer and there is no marketplace selector