The RealReal Luxury Consignment Scraper

Pull the live resale catalogue of The RealReal, the largest US marketplace for authenticated luxury consignment, as clean tabular data.

Every row is one consigned item: a single physical object, listed once, with the designer, the item class, the marketplace's own condition grade, what it originally retailed for, and what it is listed at today.

What makes this useful

The RealReal is the closest thing the resale market has to a public price book. Because each listing is a unique physical item rather than a stocked product, the catalogue is a running record of what specific pieces actually get listed at, in what condition, by which designer.

This Actor is built around that. Every row carries both prices and both derived percentages:

Field	Meaning
retailPrice	what the item cost new
listPrice	what it is consigned for now
originalPrice	the consignment price before any markdown
vsRetailPct	how far the list price sits below retail
discountPct	how far the list price sits below the was-price

So a pull for one designer, in one category, is a price-positioning table for that brand with no further work. That is the intended use: brand and merchandising teams tracking where their pieces land on the resale market, resellers pricing inventory, and appraisers checking comparables.

Two ways to browse

Fill in either or both; each entry is walked as its own catalogue.

• Designers the analytics axis. Give a list of designer names and each one is walked as its own catalogue. Pair it with Category for the designers above to compare like with like, for example every Chanel handbag.
• Category paths a whole department, for example women/handbags, men/shoes, jewelry or watches.

Or switch to url mode and paste category, designer or item links directly.

Why there is no keyword box

This marketplace does not narrow its catalogue pages by keyword. A keyword is accepted and then ignored, and the page comes back with the entire catalogue, about 3 million items, whose first results have nothing to do with what was asked for. A keyword box would therefore look like it worked while returning the wrong thing, so this Actor does not offer one and skips a pasted keyword link with an explanation rather than answering it. Browse by category or designer instead, and use the filters below to narrow.

Condition grades

The marketplace grades every item on its own five-step scale, and the grade is returned verbatim in the condition field:

Grade	Meaning
Pristine	Original condition, may include tags, dustbag or box
Excellent	Tried on or lightly worn, in great condition
Very Good	Minor signs of wear
Good	Visible wear
Fair	Heavy wear

Set the Condition grade filter to keep only one grade. When an item page is read, the inspector's own free-text note (for example "minor residue at flap") is returned separately as conditionNote.

Output fields

Returned on every row, from the result pages, at no extra cost:

itemId, sku, variantId, name, url, designer, designerId, itemClass, category, condition, conditionNote, color, gender, primaryMaterial, attributes, retailPrice, listPrice, originalPrice, currency, discountPct, vsRetailPct, isOnSale, availability, isSold, quantity, obsessionCount, isEditorsPick, badge, images, imageCount, sourceUrl, scrapedAt

Added when Read each item's page is on:

measurements (height, width, strap drop and so on), materials, description, authenticationNote, sustainability, returnPolicy, disclaimers

Added when Incremental changes for scheduled runs is on:

changeType, changedFields, firstSeenAt, lastSeenAt

Cost

Three charge events:

• Actor start, once per run.
• Item result, one per item returned. This is the main cost.
• Item page detail, one per item whose own page was read. Charged ONLY when Read each item's page is on, and only for items that actually reached the dataset. An item whose page could not be read is still returned with its result-page fields and is NOT charged this event.

Leaving the detail toggle off is the cheap mode, and it still returns the designer, condition grade, both prices and both percentages. Turn it on when you need measurements or the full item copy.

Recurring monitoring

Turn on Incremental changes for scheduled runs and the Actor remembers the previous run of the same search by itself. Later runs return only what changed, marked NEW, UPDATED or REAPPEARED.

On a consignment marketplace this is unusually informative, because each item is unique:

• an UPDATED row with listPrice in changedFields is a markdown
• an EXPIRED row usually means the item sold

Turn on Emit expired items to get those. Expiry rows are produced only after a run has completely scanned the tracked search, so a run that was capped, resumed, or stopped at the marketplace's own per-search depth limit never reports an item as gone when it simply was not reached.

Depth limit, and how to work around it

The marketplace serves at most 2000 items per search, however many match. A category like women's handbags reports hundreds of thousands of matches and still stops at 2000.

The Actor tells you in the log when a search hits that limit. To collect more, split the search:

• browse one designer at a time rather than a whole category
• narrow to a single category with the designer category field
• walk the catalogue in price bands using the minimum and maximum price fields, for example 0 to 500, then 500 to 1000, and so on

Each of those is a separate search with its own 2000-item window.

Ordering

Results come back newest first. The marketplace applies its other orderings in the browser rather than on the page this Actor reads, so no sort option is offered here rather than offering one that would quietly do nothing.

Send results into your apps (MCP connectors)

Optionally pipe results into the apps you already use, over Model Context Protocol. Authorize a connector under Apify, Settings, API & Integrations, then pick it in mcpConnectors. Notion gets a page per item, so it also needs notionParentPageUrl; other connectors get a best-effort write or a digest. maxNotifyListings caps how many items each connector receives per run.

Each connector receives a condensed summary per item, never the full record. The complete record always stays in the dataset, and leaving this empty changes nothing about the output.

Sample output

Values are placeholders, not a real listing.

{
  "itemId": "00000000",
  "sku": "EXM000000",
  "name": "Example Leather Shoulder Bag",
  "url": "https://www.therealreal.com/products/women/handbags/shoulder-bags/example-item-aaaaa",
  "designer": "Example Maison",
  "designerId": "000",
  "itemClass": ["Women", "Handbags", "Shoulder Bags"],
  "category": "Shoulder Bags",
  "condition": "Excellent",
  "conditionNote": "No visible signs of wear.",
  "color": "Black",
  "gender": "Women",
  "primaryMaterial": "Leather",
  "retailPrice": 3200.0,
  "listPrice": 1450.0,
  "originalPrice": 1700.0,
  "currency": "USD",
  "discountPct": 14.7,
  "vsRetailPct": 54.7,
  "isOnSale": true,
  "availability": "AVAILABLE",
  "isSold": false,
  "obsessionCount": 0,
  "measurements": { "Height": "8\"", "Width": "11\"", "Shoulder Strap Drop": "20\"" },
  "materials": ["Example Maison Shoulder Bag", "Black Leather", "Gold-Tone Hardware"],
  "images": ["https://product-images.therealreal.com/EXM000000_1_enlarged.jpg"],
  "detailFetched": true,
  "scrapedAt": "2026-09-11T00:00:00Z"
}

Notes

• Prices are in US dollars.
• Sold items are returned by default, because a sold item and its final price are the most useful comparable there is. Turn on Only items still available to exclude them.
• This Actor collects no personal data. Consignors are not identified anywhere in the marketplace's listing pages, and no buyer reviews exist on consignment listings, so there is no reviewer identity to collect either.
• Item pages are slower to read than result pages, so keep Max items tight when the detail toggle is on.