# Wildberries Scraper API - Парсер Wildberries и Вайлдберриз (`mattnews17/wildberries-scraper`) Actor Аналитика и парсер Вайлдберриз (ВБ): цены, скидки, реальные остатки по каждому SKU, рейтинги, продавцы с wildberries.ru. Wildberries scraper API for Russia and the largest RU marketplace — verified per-SKU stock, no duplicate billing. 100 товаров за запрос. $0.001 per product. Неофициальный. - **URL**: https://apify.com/mattnews17/wildberries-scraper.md - **Developed by:** [Matt Newsome](https://apify.com/mattnews17) (community) - **Categories:** E-commerce, Automation - **Stats:** 1 total users, 0 monthly users, 100.0% runs succeeded, 0 bookmarks - **User rating**: 4.71 out of 5 stars ## Pricing from $1.00 / 1,000 product-scrapeds This Actor is paid per event. You are not charged for the Apify platform usage, but only a fixed price for specific events. Learn more: https://docs.apify.com/platform/actors/running/actors-in-store#pay-per-event ## What's an Apify Actor? Actors are web data automations that power AI and operations. They run on the Apify platform to scrape websites, process data, connect APIs, and automate workflows. In Batch mode, an Actor accepts a well-defined JSON input, performs an action which can take anything from a few seconds to a few hours, and optionally produces a well-defined JSON output, datasets with results, or files in key-value store. In Standby mode, an Actor provides a web server which can be used as a website, API, or an MCP server. Actors are written with capital "A". ## How to integrate an Actor? If asked about integration, you help developers integrate Actors into their projects. You adapt to their stack and deliver integrations that are safe, well-documented, and production-ready. The best way to integrate Actors is as follows. - **AI agents and MCP clients** — the [Apify MCP server](https://docs.apify.com/integrations/mcp.md) at `https://mcp.apify.com` (remote, streamable HTTP, OAuth on first use). - **Agentic workflows and local Actor development** — [Agent Skills](https://apify.com/.well-known/agent-skills/index.json) with the [Apify CLI](https://docs.apify.com/cli/docs.md): `npm install -g apify-cli`, then `apify login`. - **JavaScript/TypeScript projects** — the official [JS/TS client](https://docs.apify.com/api/client/js/docs.md): `npm install apify-client`. - **Python projects** — the official [Python client](https://docs.apify.com/api/client/python/docs.md): `pip install apify-client`. - **Any other language** — the [REST API](https://docs.apify.com/api/v2.md). For usage examples, see the [API](#api) section below. For more details, see Apify documentation as [Markdown index](https://docs.apify.com/llms.txt) and [Markdown full-text](https://docs.apify.com/llms-full.txt). # README ## Wildberries Scraper — Product, Price, Discount & Stock Data **Wildberries scraper** (парсер Wildberries) that extracts structured product data from wildberries.ru, Russia's largest online marketplace. Search by keyword and get clean rows with **current price, original price, discount percentage, stock quantity, rating, review count, brand and seller** — at **$0.001 per product**, and never billed twice for the same SKU. > **Unofficial tool.** This Actor is not affiliated with, endorsed by, or > connected to Wildberries or its operators in any way. "Wildberries" is used > solely to describe the publicly accessible website this Actor reads. All > trademarks belong to their respective owners. Cheap by design: **100 products per request**, straight from JSON. No headless browser, no per-product page fetches. Throughput is deliberately throttled rather than fast. Wildberries returns HTTP 429 to parallel requests from datacenter IPs, so the defaults are 1 concurrent request at 6 per minute — about **600 products per minute**, so a 1,000-product run takes roughly two minutes. Raising those defaults on datacenter IPs does not make it faster, it makes it return nothing. With residential proxies you can raise both. ### What you get | Field | Description | |---|---| | `sku` | Wildberries product ID | | `title` | Product name | | `brand` | Brand name (`null` when the seller left it blank) | | `price` | Current selling price | | `originalPrice` | Pre-discount list price, when genuinely marked down | | `discountPct` | Discount percentage, computed from the two prices | | `currency` | `RUB` | | `inStock` | Availability as a boolean | | `stockQuantity` | Units available — not just in/out | | `rating` | Average review score | | `reviewCount` | Number of reviews | | `supplier` | Selling entity | | `url` | Canonical product URL | | `scrapedAt` | ISO timestamp | `categoryPath`, `image` and `description` are always `null` in Wildberries mode — the search endpoint does not return them. Image URLs are deliberately omitted rather than guessed: Wildberries shards images across numbered CDN baskets whose ID ranges shift over time, and a plausible-looking dead link is worse than an honest null. ### Real per-SKU stock (optional) Most Wildberries scrapers report the search endpoint's quantity field as if it were stock. It isn't. Measured across three runs (100 headphones, 200 smartphones, 200 sneakers) it returns **one repeated value for 65–95% of rows** — 46 in one run, 48 in two others — with only a handful of genuine numbers. A "stock" column that says the same thing about two thirds of the catalogue cannot tell you anything about any single product. Switch on **Verify real stock per product** and each search page makes one extra batched request to Wildberries' product-card endpoint, which does hold real per-SKU quantities. Measured on the same 100 headphones: | | Search field | Card endpoint | |---|---|---| | Distinct values per 100 rows | 7 | **22** | | Distinct values per 300 rows | — | **42** | | Genuine zero-stock rows found | 0 | **10 per 100** | | Low-stock numbers (1, 2, 13, 14…) | almost none | **yes** | So out-of-stock detection and low-stock alerts actually work — which is the question sellers are really asking. **The ceiling, stated plainly.** Wildberries clusters a large share of rows on a single high value, and **that value moves between runs**. Measured: one run topped out at 46 with 61 of 100 rows on it; a larger run topped out at 52 with 157 of 300 rows on it — same region, same kind of query. Below the ceiling the numbers vary genuinely (42 distinct values in that 300-row run) and include real zeros. So treat a row sitting on the run's maximum as **"at least this many"**, and anything below it as a real count. Every run's log prints the ceiling it actually observed, e.g.: ``` stock ceiling observed this run: 52. Wildberries clusters a large share of rows on the run maximum, and that maximum moves between runs (46 and 52 both seen)... ``` There is deliberately **no** hardcoded cap and no per-row "capped" flag. An earlier version asserted the ceiling was always 46; on the very next larger run that mislabelled 161 rows. Reporting the observed maximum per run cannot go stale in that way. Every row also carries `stockSource`: | Value | Meaning | |---|---| | `card` | Quantity verified against the product card | | `search-fallback` | Card lookup was requested but did not cover this SKU — the search value is kept, unverified | | `search` | Verification was off; this is the unreliable listing value | **Cost.** One `fetch-detail` event per batch of 50 products, not per product — so verifying 1,000 products costs 20 extra events, and the per-product price is unchanged. Off by default, because you should not pay for it unless you want it. ### Sample output Real rows from a live run, unedited — search term `наушники`, 100 products, stock verification on: | title | brand | price | was | disc % | stock | src | rating | reviews | |---|---|---|---|---|---|---|---|---| | Наушники проводные совместимы… | STIMAXON | 358 | 960 | 62.7 | 52 | `card` | 4.5 | 14325 | | Наушники проводные Type-C совм… | `null` | 296 | 2100 | 85.9 | 2 | `card` | 4.9 | 3359 | | Наушники проводные | Borofone | 203 | 798 | 74.6 | 0 | `card` | 4.7 | 617 | Those three rows are chosen to show the range: one sitting on the run's ceiling (52, read as "at least 52"), one genuinely down to its last 2 units, and one actually out of stock. That is the difference the card lookup makes — without it all three would have read the same placeholder. That run: **100 rows, 100 unique SKUs, 0 duplicates, 0 invalid discounts, 100/100 relevant to the query**, 24 distinct stock values, 9 genuine zeros. Every field populated except `brand`, which 25 sellers left blank at source. One complete row, every field the Actor emits: ```json { "sku": "337343445", "title": "Наушники проводные совместимы с iPhone с микрофоном", "brand": "STIMAXON", "price": 358, "originalPrice": 960, "discountPct": 62.7, "currency": "RUB", "inStock": true, "stockQuantity": 52, "stockSource": "card", "rating": 4.5, "reviewCount": 14325, "supplier": "STIMAXON", "categoryPath": null, "url": "https://www.wildberries.ru/catalog/337343445/detail.aspx", "marketplace": "wildberries.ru", "country": "russia", "image": null, "description": null, "extractedVia": "wb-api", "scrapedAt": "2026-08-04T17:05:41.092Z" } ``` Note the discount depths. Wildberries sellers routinely inflate the list price, so a 95% markdown is the source's number, not a real saving. The Actor reports both prices and computes the delta; judging whether the "was" price was ever real is left to you. ### Use cases - **Daily price monitoring** — track competitor pricing and react to changes - **Discount and promotion tracking** — real markdown depth, since both list and sale price are returned - **Stock and low-stock monitoring** — with stock verification on, real per-SKU quantities let you spot a competitor down to their last few units, and genuine out-of-stock rows instead of a placeholder - **Repricing feeds** — push live market prices into your own pricing engine - **Brand and seller monitoring** — watch who is selling your brand and at what price - **Assortment research** — see what sells in a category and at what price points ### Input Runs with **zero configuration** — press Start for a sample run. ```json { "site": "wildberries", "searchTerms": ["наушники"], "maxItems": 1000 } ``` Russian keywords give the best coverage. Each term is paginated until results run out or `maxItems` is reached. `maxItems` is a hard cap, so your bill stays predictable. ### Reliability Most scrapers break because they hardcode CSS class names that die on the next redesign. This one reads a JSON endpoint and validates every field on the way out: - **You are never billed for duplicate rows.** Wildberries' own pagination repeats products across pages, and overlapping search terms surface the same items. Rows are deduplicated by SKU within a run before anything is emitted or charged. If you ask for 1,000 products and a keyword only has 600 unique ones, you get 600 real rows — not 1,000 rows with 400 repeats. - Items that do not parse are **skipped, never emitted half-filled** - A drop in the map-success rate raises an explicit warning naming the likely cause - Rate-limit responses (429) are retried against fresh sessions automatically - Every run reports its **yield percentage**, so a partially-degraded run is visible instead of silently returning half the data - **Fabricated responses are rejected, not sold to you.** Wildberries does not only answer 429. A flagged request can also receive valid-looking JSON containing invented products — observed live: a search for `наушники` (headphones) that came back carrying a single evening dress while the envelope claimed 100 results. Such a payload is discarded and retried against a fresh session, so no decoy row is ever emitted or charged. Every run reports `decoysRejected=N`. - **Anti-bot wrappers are recovered, not thrown away.** Some valid responses arrive wrapped in an injected `