South Asia's biggest marketplace, as four catalogues in four currencies

Four storefronts: Pakistan in PKR, Bangladesh in BDT, Sri Lanka in LKR and Nepal in NPR. They are genuinely separate catalogues rather than one catalogue in four skins. On the same query the first five product ids are entirely disjoint across all four, prices differ by magnitudes that are not a currency conversion of one shared number, and the match counts differ too - a search for shoes returns 1,881 matches in Bangladesh while Pakistan, Sri Lanka and Nepal all reach Daraz's display ceiling. Daraz's Myanmar storefront, daraz.com.mm, is not offered: as of 23 September 2026 the domain no longer serves a working site at all, and asking for it returns a clear MARKET_UNAVAILABLE rather than a confusing error.

Because three of the four storefronts print the same symbol for three different currencies. Pakistan, Sri Lanka and Nepal all show prices as Rs., meaning Pakistani, Sri Lankan and Nepalese rupees, which are not the same unit of money, and Daraz publishes no currency code anywhere in its data. If the API returned a plain price field, a caller pulling two countries into one table would silently be adding together three different currencies. So the field is price_pkr, price_bdt, price_lkr or price_npr, the currency is always stated in the response, and the mistake becomes impossible to make by accident.

Forty products a page. Daraz reports an exact match count up to 4,080 and truncates its own figure there, so the response tells you which case you are in: total_results carries Daraz's number, total_results_capped says whether that number is the real count or the ceiling, and last_page plus no_more_pages give you the real end. A search for Samsung Galaxy S23 Ultra 512GB returns 4 matches and exactly 4 rows; a search for Nike Air Force 1 returns 3,164 and ends at page 80, which is 3,164 divided by 40. Asking for a page beyond the end returns an empty page with an explanation in page_note rather than filler, because past its real last page Daraz keeps answering with forty unrelated products and nothing in its response says they are not results.

Only the ones measured to change the result, and the page documents what each one does to the count. A price band cuts the query phone from 4,080 matches to 311 with all 40 returned rows inside the band, and a category narrows the same query to 2,108. A minimum rating and a shipping region genuinely change the rows - the rating filter removed every row below four stars where the unfiltered page had two, and the region filter returned 40 of 40 rows from Punjab against an unfiltered split of 15 Punjab and 25 Sindh - but Daraz does not recalculate its reported total for either, so on those two the rows are authoritative and the total is not, and the API says so in the parameter description rather than letting you find out. Brand, seller and service filters are deliberately not offered: Daraz advertises them, accepts them and then ignores them, and a filter that silently does nothing is worse than no filter. Brand and seller are still returned on every row, so you can filter on them yourself.

You can sort by price, on page one. Cheapest-first and dearest-first both work and were checked row by row: 39 of 39 adjacent pairs came back correctly ordered in each direction. But Daraz returns the identical forty rows on page one, page two and page ten of a price-sorted query, so paging a price sort would hand you the same products over and over while looking like progress. Asking for a page beyond the first together with a price sort returns INVALID_PARAM explaining exactly that. To walk a catalogue by price, drop the sort and step through price bands instead.

The full description as clean text and as the original HTML, every gallery image and video, the colour and size variant axes with each value's own image, the SKU table with per-order limits, the category path, the brand link, availability and the return policy. Across 18 products in six categories, title, description, brand, images, category path and currency were present on all 18, and the variant axes on 16 - the two without them are genuinely single-variant items. The product record also carries the charged price, rating, review count, units sold and seller name, which Daraz does not put on its product page at all; the API resolves them for you in the same call and tags them with price_source so you always know which of Daraz's own documents a number came from.

It depends on the country, and the honest answer is published rather than averaged away. On Pakistan and Sri Lanka, rating arrives on roughly 30 to 34 rows in 40 and units sold on 21 to 33. On Nepal it is far sparser - on one 40-row laptop search, rating arrived on 2 rows and units sold on 1 - because the Nepali catalogue carries much less review activity. If you are building on rating or sold count, choose the country with that in mind. Daraz states no rating scale anywhere, so rating_scale is returned as null rather than being assumed to be out of five, and any field Daraz does not publish comes back null rather than as a guess.

Up to 50 reviews a page and 1,000 deep per product, and the aggregate figures cover the whole product however deep you page. The star filter and Daraz's own histogram agree exactly: filtering to one star returned 727 where the histogram says 727, and to five stars 6,300 where it says 6,300. One thing worth knowing is that the histogram does not add up to the stated total, and that is Daraz's own arithmetic rather than a rounding error on our side - on one product the five star buckets sum to 8,308 against a stated 8,362, because Daraz withholds some reviews from the public list. The count it withholds is returned as hidden_review_count so you can see the gap instead of discovering it as a bug.

Daraz API is a ReefAPI endpoint group for south asia's biggest marketplace across pakistan, bangladesh, sri lanka and nepal: four catalogues, four currencies, with the charged price and the strikethrough kept apart. It returns live JSON through POST requests under /daraz/v1.

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

No login to Daraz 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.

Daraz actions currently cost 2-3 credits per successful call. Failed or blocked calls are free. All APIs draw from one credit pool.

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