Search Tokopedia and read Indonesian marketplace prices as JSON

Because a bare numeric id is not addressable on Tokopedia, and detail no longer pretends otherwise: product_id has been removed from that action's schema, so a call without a url now answers MISSING_PARAM naming url, immediately and without going upstream. Take the url straight from the search row; you can drop everything from the ? onward, since the query string is only Tokopedia's click tracking. reviews is the exception and still accepts product_id, because there it works.

It is Tokopedia's units-sold badge, and on a search row it is text, not a number. Measured values include "8 terjual", "29 terjual" and "50+ terjual" - terjual is Indonesian for sold, and the plus sign means Tokopedia has stopped reporting the exact figure and banded it. The detail action gives you the same product as a plain numeric string ("8"), so if you need arithmetic, read it there, or parse the leading digits and treat a trailing plus as a floor rather than an exact count.

price_value is a whole-rupiah integer - 109000 for Rp109.000, 1423056 for Rp1.423.056 - because the rupiah has no practical subunit in retail pricing. The companion price field is the display string Tokopedia prints, where the dot is a thousands separator, not a decimal point. Parsing "Rp1.735.000" as a float in a locale that reads dots as decimals gives you 1.735 rupiah. Use price_value for anything numeric.

Tokopedia's review widget and the counter on the product page are fed separately, and the widget can return nothing for a product whose counter is non-zero. A measured call against a product whose detail reported review_count "3" returned ok:true, meta.total 0 and an empty reviews list, with stop_reason "complete" - so it is a real empty result, not a truncated page. Treat review_count as a page-level hint and meta.total on the reviews action as the number you can actually fetch.

You get NOT_FOUND with retryable false and a message saying the product was removed or never existed. Tokopedia's search index keeps serving delisted listings for a while, so a url that came out of a fresh search can already be dead - three of the first four best-selling results for one keyword behaved that way. That answer used to arrive as a retryable upstream error, which meant a sync could keep re-buying a permanently negative result; it is now a closed verdict. Delete the row rather than queueing it.

20 per page, with page running from 1 to 100, so 2,000 rows is the ceiling per query no matter how large meta.total is - "laptop gaming" reported total 38,583. meta.has_more and meta.next_page tell you when to stop. To go deeper, narrow with price_min / price_max, condition or rating and run several queries, or use shop_products, which walks a single seller's catalog exhaustively.

They are lifetime store totals, not current inventory. A measured profile returned active_product_count 95 (what is listed right now) against product_sold 597,775 and transaction_success 287,109 (both since the shop opened), plus total_favorite 36,072, which is the follower count rather than a like count. open_since is an English month-and-year string ("July 2014"), location / city / district split the address, and status is a numeric code (1 on an open shop) alongside the is_open boolean.

Two fields, and only one is machine-readable. date is the Indonesian relative phrase Tokopedia prints - "2 minggu lalu" (2 weeks ago), "3 bulan lalu" (3 months ago) - while date_raw is a Unix epoch in seconds delivered as a string ("1785992410"). Sort and store on date_raw. Also expect author, author_id and author_url to be null whenever is_anonymous is true, which was the case for every review measured; shop_reply carries the seller's canned response when there is one.

Tokopedia API is a ReefAPI endpoint group for tokopedia It returns live JSON through POST requests under /tokopedia/v1.

Yes. ReefAPI starts with 1,000 free credits, no card required. Tokopedia calls use the same shared credit balance as every other ReefAPI engine.

No login to Tokopedia is needed for the API response. You call ReefAPI with your x-api-key header, and the playground can run live examples before you create a production key.

The page example is captured from a live search call, and production requests fetch live data through ReefAPI rather than a static sample.

Tokopedia actions currently cost 1 credit per successful call. Failed or blocked calls are free, and all APIs draw from one credit pool.

Yes. Connect ReefAPI once through MCP and your assistant can call tokopedia actions with the same key, credit pool and JSON envelope used by normal REST requests.