TikTok Shop US Scraper — Products, Prices & Stock
Pricing
from $1.00 / 1,000 product snapshots
TikTok Shop US Scraper — Products, Prices & Stock
Extract public TikTok Shop US product prices, SKUs, stock, ratings and sold counters by keyword or product URL. JSON/CSV. $1 per 1,000 snapshots plus Apify platform usage. Keyword first page only.
Pricing
from $1.00 / 1,000 product snapshots
Rating
0.0
(0)
Developer
Coor Yu
Maintained by CommunityActor stats
0
Bookmarked
3
Total users
2
Monthly active users
3 days ago
Last modified
Categories
Features
Share
Public beta — US$1 per 1,000 product snapshots, plus Apify platform usage. No Actor-start fee.
Read public TikTok Shop US product pages and the first public keyword-results page. Export structured product snapshots for product research, price comparisons and downstream analysis. No login, paid analytics source, automatic paid proxy or browser is used.
What it does
- Discover products by keyword on the first public server-rendered results page, source-dependent and limited to at most 30 candidates per keyword.
- Read direct US product URLs, including supported
www.tiktok.com/shop/pdp/...links. - Extract the product ID, title, description, current/original price, currency, SKU prices and stock, images, shop name, public rating/review count, shipping and the public sold counter when available.
- Deduplicate by product ID across all inputs within one run. Direct URLs are processed first.
- Optionally include up to five embedded review samples. These are not all reviews and are not guaranteed to be the newest.
- Compare a supplied earlier
SNAPSHOTagainst the current snapshot for compatible price, stock and counter changes. - Produce product-only datasets, JSON, spreadsheet-safe CSV, and separate diagnostics.
Honest limits
- US storefront only. Other countries, profiles, videos, arbitrary websites, shortened links, categories and shop URLs are not supported inputs in this release.
- First keyword page only. There is no cursor/API pagination. If TikTok reports more results, the run emits
FIRST_PAGE_ONLYinNOTICES. Increasing a limit does not invent more products. - Counter scope is explicit. A product may display global lifetime items sold, excluding returns. That is not US-only sales, not a 7/30-day metric and not official GMV. Rounded counters are marked. No purchased GMV or revenue split is provided.
- Missing means null, not zero. SKU stock may be incomplete, stale or hidden. A stock sum is returned only if every returned SKU has a usable stock value. Listing-only results do not pretend to contain SKU stock.
- No silent paid fallback. A denied/rate-limited response is reported. No proxy, unblocking service, vendor API or other Actor is purchased automatically. Cloud IPs can behave differently from the local development environment.
- No full review scraping. Embedded samples are optional, stay inside the product row and omit reviewer names, avatars and profile links.
- No built-in recurring service.
previousProductsis a caller-supplied earlier snapshot. This Actor does not schedule itself or maintain a cross-customer historical database. - Search relevance is not guaranteed by TikTok. The default title filter requires each case-insensitive Unicode word/number token as a substring. Disable it to accept broader source results.
Quick start
Keyword discovery with detail pages:
{"keywords": ["air fryer"],"maxProducts": 10,"maxProductsPerKeyword": 10,"includeProductDetails": true,"strictKeywordMatch": true,"maxRequests": 20}
Direct public product URL:
{"startUrls": [{ "url": "https://shop.tiktok.com/us/pdp/cat-nail-clipper-by-potaroma-adjustable-sizes-built-in-file-safe-for-kittens-cats/1731578642912612516" }],"maxProducts": 1,"includeReviewSamples": false,"maxRequests": 3,"maxRetries": 0}
The example product is a live public listing, not a guaranteed permanent test fixture. It can change or disappear.
Main input controls
| Input | Default | Meaning |
|---|---|---|
keywords | [] | Up to 20 search phrases, 150 characters each |
startUrls | [] | Up to 100 public US product URLs |
maxProducts | 20 | Global maximum unique product results, 1–500 |
maxProductsPerKeyword | 25 | Per-keyword result upper bound, 1–30 |
includeProductDetails | true | Open discovered products; additional HTTP work |
strictKeywordMatch | true | Require the keyword's tokens in the title |
includeReviewSamples | false | Include already-embedded review samples only |
minPrice, maxPrice | not set | Current US price filter |
minSoldCount, minRating | not set | Public-counter and product-rating filters |
previousProducts | [] | Earlier SNAPSHOT array for comparison |
maxRequests | 100 | All HTTP attempts, including redirects/retries |
maxRunSeconds | 180 | Application runtime guard, checked between requests |
requestTimeoutSecs | 20 | In-flight request timeout, 3–45 seconds |
maxRetries | 1 | Only transient network/timeouts and selected 5xx statuses |
requestDelayMillis | 250 | Delay between public requests |
The application runtime guard is not a precise invoice cap. An in-flight request can take up to the request timeout beyond it. Also set the Apify run timeout and appropriate memory.
Output
The default dataset contains only valid product rows, never error/notice objects. Important fields:
productId(string; large IDs are never converted into imprecise numbers),url,country,title,description.price,originalPrice,currency,discountPercent,variants,stockQuantity,stockIsComplete,inStock.soldCount,soldCountRaw,soldCountIsApproximate,soldCountScope,soldCountNote.rating,reviewCount,shopId,shopName,shopRating,shipping,categories,specifications.dataSource,dataLevel,sourceType,sourceInput,position,capturedAt,missingFields,warnings,changes.
Key-value records:
| Record | Contents |
|---|---|
OUTPUT | Full product array |
OUTPUT.csv | UTF-8 BOM CSV; formula-like untrusted text is neutralized |
SUMMARY | Result counts, request counts, source coverage, errors and stop reason |
ERRORS | Invalid inputs, denied/unavailable pages and parsing failures |
NOTICES | First-page-only and no-result notices |
SNAPSHOT | Minimal array suitable as a later run's previousProducts |
Zero results due exclusively to source/parsing errors make the cloud run fail after saving diagnostics. A legitimate empty keyword feed or all-filtered results are reported as empty, not silently replaced with recommendations.
Snapshot comparisons
Copy an earlier run's SNAPSHOT array into previousProducts of a later run. Price deltas require compatible known currencies. Stock deltas require known quantities. Counter-difference estimates require matching known counter scope, exact counters and a non-decreasing counter. Future timestamps and counter resets do not produce fake sales figures.
Counter differences can still be affected by returns, corrections and source timing. They are not verified net sales, not historical GMV and not a generated 7/30-day series.
Pricing and spending controls
Price: product events plus platform usage
- US$0.001 per unique product snapshot saved to the dataset (US$1 per 1,000), at the same service rate for every plan tier.
- Apify compute, storage, traffic and other platform usage are charged separately to the caller. This is not an all-inclusive US$1 bill. The Store Pricing tab and the run's usage detail show the applicable charges.
- No Actor-start fee. There are no separate events for SKU arrays, images, shipping fields or embedded review samples.
- Listing-only and detail snapshots have the same product-event price. Missing optional fields are explicitly null; a result is not a promise that every source field exists.
- Errors, rejected inputs, notices and filtered-out products do not produce a product event. Platform resources used while attempting a request can still be charged, including when no results are returned.
Example: 3 delivered products incur US$0.003 in product events, plus the actual platform usage. A user spending limit for events is not a combined cap on the separately billed platform usage; set a platform timeout, memory and the input's HTTP/time limits as well.
The minimum permitted event budget is US$0.01. It is not a minimum spend: only delivered product events are charged. The default event budget is US$0.05 and can stop a run before its requested product maximum; increase it explicitly when needed.
The Actor uses the custom product event through the SDK's push/charge integration and refuses configurations with both product and apify-default-dataset-item, preventing duplicate result-event configuration. It checks event affordability before further network work. Diagnostics are kept outside the product dataset.
Dataset reconciliation on a resurrected run reduces duplicate outputs; SDK retries handle event API idempotency. This does not constitute a guarantee of exactly-once behavior across every possible distributed-system crash boundary.
Local development and verification
Requires Node.js 22 or newer.
npm ci --ignore-scripts --no-audit --no-fundnpm run checknpm testnpm run smoke:fixtures
smoke:fixtures reads minimized captured public responses. It makes no external network call and no Apify call. Unit-test mutations/mocks are synthetic and are not presented as live results.
$npm run smoke:live
smoke:live makes a bounded local public-HTTP check (no SDK initialization and no paid provider). It writes fresh results to generated/reports/live. It is not a cloud-run or billing verification.
Deployment and verification
- Source version:
0.1; package version:0.1.0; release tag:latest. - Memory bounds: 256–512 MB; no Chromium image or automatically purchased proxy.
- The input, dataset, output and Actor definitions pass Apify's official schema validators.
- The parsing, filtering, request-budget and billing-adapter unit tests pass on Node 22.
- A real Apify datacenter run returned three detail products with four public HTTP requests and no errors. This is a small-sample check, not a claim of universal or long-term success.
Check the active Store Pricing page for the current product-event rate and the separate platform-usage model. Do not rely on an old source archive as proof of current pricing.
Data use and support
Read only publicly accessible pages; do not supply login cookies or credentials. Ensure your use, retention and redistribution comply with applicable law, TikTok's terms and any contractual obligations. Public access is not a blanket license for all downstream uses. Report a problem with the run ID and ERRORS/SUMMARY records; remove private information before sharing logs.