Vinted Multi-Country Scraper

Scrapes Vinted, Europe's largest C2C second-hand fashion marketplace, across its 27 country marketplaces: keyword search with brand, price and condition filters, item and seller pages, and a seller's full member feedback history.

What you get

• Search mode: keyword search on any country's catalogue with source-side brand, category, price and condition filters. Every listing carries price + local currency, brand, size, condition, view count, favourite count, photos, and the seller reference.
• Item detail (toggle, on by default): full description, colour, upload date and the schema.org condition.
• Seller profiles (toggle): the seller's city, country, feedback reputation with positive/neutral/negative counts, item and follower counts, bundle discount, response time and online status.
• Reviews mode: every feedback entry for a seller as its own row (rating, comment text, date, reviewer, linked item), plus the seller's summary rating on each row.
• URL mode: paste item pages, seller/member pages, or catalogue/search pages (a page= parameter starts the walk at that page). Each URL's own hostname picks its market, so links from different countries can be mixed in one run.

Quick start

1. Pick a Mode.
2. Search mode: type Search keywords (e.g. nike air max), optionally narrow with Category, Brand, Condition or price. URL mode: paste item or member links. Reviews mode: paste member ids or links.
3. Pick the Country marketplace (defaults to Germany). Prices come back in that market's own currency (EUR, PLN, GBP, SEK, ...).
4. Set Max items if you want a bounded run. 0 (the default) keeps going until the source runs out of results.

Output

Field	Sample value
itemId	"0000000000"
title	"Sample knitted winter sweater"
brandTitle	"Sample Brand"
priceAmount	12.5
currency	"EUR"
condition	"Very good"
size	"M"
color	"Green"
viewCount	142
favouriteCount	18
seller.username	"sample_seller"
seller.feedbackReputation	4.9
url	"https://www.vinted.de/items/0000000000"

Samples above are illustrative placeholders, not real listings. A review row (Reviews mode) carries reviewId, rating, body, createdAt, reviewerUsername and the seller's summary rating.

The second, larger price the site displays next to the listing price is the buyer-protection fee added at checkout, not a former price - this actor therefore does not emit a "was price" field.

Modes and scope (what applies where)

Input	Search	URL	Reviews
Country marketplace	yes	fallback only	fallback only
Keywords, category, brand, condition, price, sort	yes	no	no
URLs (item / member / catalogue pages)	no	yes	no
Sellers	no	no	yes
Fetch item detail / Fetch seller profiles	yes	yes	n/a (reviews are the product)

In URL and Reviews modes each pasted URL always uses its own market (the marketplace keeps one catalogue per country, and a seller's full item list only shows on the seller's home market). The Country marketplace setting is the fallback for bare numeric ids and unknown links.

Notes and limits

• Anonymous search honours two sort orders: the default relevance ranking and "newest first". Price sorting is only offered to signed-in members on the site, so it is not offered here.
• Condition labels come back in the market's language (e.g. "Sehr gut" on the German market); the schema.org condition enum (conditionSchema, e.g. UsedConditions) is language-stable.
• Sold and removed items 404 on read. The run reports them as unavailable and continues; in incremental mode a tracked listing that disappears after a complete scan is reported as EXPIRED.
• The actor fails loudly when the marketplace refuses to serve pages (a connection problem is never reported as "no results"); when some pages are refused but real rows were delivered, the run reports Partial results:.

Resume & recurring updates

Two different tools, both optional:

• resumeFromRunId: continue ONE interrupted run. Paste a previous run id or dataset id and the run skips (and does not re-charge) records that run already collected.
• incrementalMode: recurring monitoring of the same search/URLs/sellers (e.g. daily). The first run returns everything as NEW; later runs return only NEW, UPDATED and REAPPEARED records, and listings that sold or were removed come back as EXPIRED after a complete scan. State is kept separately per mode, country and setup; name a campaign with stateKey to keep it stable or deliberately share it. emitUnchanged and emitExpired return extra rows you already have (and bill for them), so both default to off.

Billing

Pay per event: a start event, a per-result event for every listing or review returned, and a detail-enrichment surcharge charged once per listing whose detail page and/or seller profile was actually fetched and returned. Suppressed (unchanged) incremental rows are never charged the surcharge.

Send results into your apps (MCP connectors)

Select one or more MCP connectors under mcpConnectors to mirror results into the app you already use, alongside the dataset. Supported: Notion, Linear, Airtable, and Apify. Authorize a connector once under Apify -> Settings -> Integrations.

• mcpConnectors: the connectors to export into. The connector receives a condensed, human-readable summary per item (title + key fields), not the full JSON; the complete record always stays in the dataset. Leave empty to skip.
• notionParentPageUrl: for the Notion connector, the URL (or id) of the page under which item pages are created. Required for Notion; ignored by other connectors.
• maxNotifyListings: cap on items written to each connector per run (default 50). Does not affect the dataset.

Computing usage

A search of 48 listings per page costs one page read per result page. A bounded run (Max items: 24, detail + seller profiles on) reads one result page, then 24 listing pages and 24 seller profiles. Local development runs honour APIFY_PURGE_ON_START=0 and the platform injects the proxy settings itself.