Whatnot Scraper — Listings, Sellers, Reviews & Live Chat

Collect public Whatnot listings, seller profiles, reviews, shows, category feeds and seller rankings. Observe active auction bids and live chat within a chosen time window. Use results for product research, seller comparisons and listing change tracking.

No Whatnot login, cookies or API key is required. Results include source references and coverage information, so you can distinguish a complete requested scope from a capped or interrupted collection.

Quick start

For a small product search, use:

{
  "modes": ["search"],
  "keywords": ["pokemon cards"],
  "searchVertical": "PRODUCT",
  "maxPages": 1,
  "maxItems": 24,
  "maxRequests": 10,
  "maxRunSeconds": 60
}

The default input collects one category's public activity totals. Start with a small limit, inspect the Dataset and SUMMARY, then increase the scope. Several modules can be combined in one run.

Available data

Module	Targets	Results
search	keywords or searchUrls	Listings, shows, sellers or categories, depending on searchVertical
shop	sellerUsernames, optional shopQuery	Seller shop listings
listings	listingUrls	Direct public listing details
seller	sellerUsernames	Followers, following, lifetime sold count, ratings and available profile details
reviews	sellerUsernames	Written reviews, dates, rating dimensions and seller responses; reviewer identities are excluded
shows	sellerUsernames	Accessible shows, status, scheduled time, viewers, watchlists, images and categories
category	categorySlugs	Exact category/tag activity totals
categories	No target required	Public category catalog IDs, labels and types
categoryShows	categorySlugs	Public category show feeds
categoryRankings	categorySlugs, rankingInterval	Source seller leaderboard for the category and selected period
auctionSnapshot	liveUrls	At most one observed active auction per show: bid, currency, bid count and end time
chat	liveUrls	Public live comments, message IDs, available public author details and role flags

Listing fields include IDs, title, description, price/currency, quantity, availability, sale format, images, seller and category. Source-provided extras include catalog price ranges, creation/update times, offer availability and current bids. Optional fields stay null when unavailable.

Search and filters

searchVertical accepts PRODUCT, LIVESTREAM, USER, CATEGORY or UNIVERSAL. UNIVERSAL combines four separate bounded searches; its order does not reproduce the website's blended Top results.

searchUrls accepts public Whatnot /search links with query, searchVertical and supported native JSON sort values. Unsupported filtering parameters are rejected; tracking parameters are removed.

listingOrder supports bestMatch, priceAsc, priceDesc and newest. Product filters include saleFormats, categoryIds, liveOnly, minPrice, maxPrice and priceCurrency. Filters apply locally after fetching pages, so a page cap can leave matches undiscovered. There is no currency conversion. Product filters apply only to product rows in UNIVERSAL searches.

Reviews, rankings and live data

Reviews default to the last 30 days. reviewDays=0 requests the accessible connection without a date cutoff, still bounded by limits. Written reviews do not necessarily equal the profile's rating count, which can include star-only ratings. Complete review history is not guaranteed.

Rankings use Whatnot's public seller leaderboard and its returned order. rankingInterval is CURRENT_DAY, CURRENT_WEEK or LIVE. Lifetime sold counts beside ranks are not sales during that period. A live leaderboard may be empty.

For auctions or chat, supply current public live-show URLs: https://www.whatnot.com/live/<show-uuid>. Up to three distinct shows are supported. auctionWaitSeconds is 5–30 seconds. Zero auction rows mean no eligible event was observed in that window. Observed bids and asking prices are not finalized sale prices or paid orders.

chatDurationSeconds is 5–600 seconds. includeRecentMessages controls the initial recent batch. Reconnect recovery batches are marked and deduplicated within the run. Chat is best effort: no complete historical transcript, guaranteed lossless recovery or deduplication across independent runs. Increase both the platform timeout and maxRunSeconds for longer captures; the platform deadline takes precedence.

Listing change monitoring

Set monitorChanges=true for PRODUCT search, shop or direct listings. The first scan emits listingChange initial snapshots. Later scans emit newly observed listings and changes to the established business fields. Unchanged observations produce no rows. Missing items in partial scans are never labeled sold or removed. Catalog, offer and bid enrichment fields are snapshots rather than additional change triggers.

Keep the same monitorNamespace for a baseline; choose another for an independent baseline. State is stored in your account and bounded to 10,000 listing IDs and 8 MiB per scope. Concurrent runs of the same scope report MONITOR_SCOPE_BUSY; interrupted locks expire within 15 minutes.

For an unresolved write from an older run, select its Dataset in recoveryDatasetId to permit read-only reconciliation, or resurrect the original run. Without access, recovery stops with MONITOR_RECOVERY_DATASET_READ_REQUIRED rather than replaying uncertain results.

Output and coverage

Dataset rows contain recordType, recordKey, target, data, completeness, coverage, provenance, retrievedAt and schemaVersion. data holds the business fields. Record keys deduplicate within a run. Receipt times are distinguished from source-provided dates; missing source values remain null.

The Output tab links to:

• Dataset: primary JSON records with standard Apify exports.
• SUMMARY: actual saved count, target outcomes, pagination, limits and errors.
• EXPORTS: CSV files grouped by record type, with common business fields flattened. These copies are not charged again. Oversized or timed-out exports report an error while retaining the Dataset.

fullCoverage=true means the requested accessible scope passed its checks, not all Whatnot data or all history. Page caps, source restrictions, dynamic pagination, item limits and spending limits can produce partial results. Consult SUMMARY for the final outcome. A valid empty source connection may succeed with zero rows.

Sample output

Shortened category result observed on 2026-10-13. Counts change over time; complete records also include keys, coverage and provenance.

{
  "recordType": "categoryInsight",
  "target": "fragrances",
  "data": {
    "category_id": "TGl2ZXN0cmVhbVRhZ05vZGU6NTQz",
    "slug": "fragrances",
    "name": "Perfume & Cologne",
    "viewer_count": 773,
    "livestream_count": 19,
    "metric_scope": "category_total",
    "observed_at": "2026-10-13T10:18:34.763952+00:00"
  },
  "completeness": "complete",
  "retrievedAt": "2026-10-13T10:18:34.763952+00:00"
}

Pricing

The configured price is $1.50 per 1,000 saved result rows ($0.0015 per row), plus the Actor Start event at $0.00005 per start up to 1 GB of allocated memory. The Store pricing panel is authoritative for current prices and any discounts.

One saved row is one result, including a listing change, ranking entry or live comment. Errors and coverage reports in SUMMARY, and CSV copies, do not create additional result charges. Empty or unchanged scans can still incur the start event. Use the platform's maximum cost per run and maxItems to control spending. Collection stops at the result spending limit.

Limits and troubleshooting

Limits are 100 targets of each kind, 100 pages per connection, 10,000 saved items, 1,000 source requests/connection attempts, and 600 processing seconds. The default platform timeout is 120 seconds. There is no unlimited mode.

• Partial output: inspect SUMMARY and increase only the relevant page, item, request or time limit.
• No auction/chat output: check that the show is live and public and that the observation window has enough activity.
• Restricted source: HTTP 401, 403 and 429 stop that access path. The Actor does not bypass challenges or replay logged-in credentials.
• Temporary failures: network errors and selected server errors receive at most two bounded retries within the request budget.
• Unknown fields: null means the public source did not provide a verified value.

This Python Actor uses the Apify SDK, public Whatnot pages and read-only public query/live interfaces. Upstream interfaces can change. It does not retrieve private orders, payment details, shipping credentials or buyer identities from reviews. It is independent and not affiliated with Whatnot.

For help, open an issue in this Actor's Issues tab with the run ID, input and expected outcome. Do not include credentials or private account data.