Walmart Intelligence Scraper

What does Walmart Search Scraper do?

Walmart Search Scraper collects source-backed public product, category, product-detail, best-seller, deal, and review records from Walmart.com or Walmart.ca. Search mode is the default. The Actor preserves market-specific prices and currencies, seller and availability fields, ratings, review counts, images, and source provenance. Source issues are reported in run logs and the OUTPUT metadata, not as dataset rows. It does not sign in, solve CAPTCHA, interact with challenges, bypass access controls, or invent product data.

Why scrape Walmart search results?

Use the Actor for public catalog research, product discovery, price and availability snapshots, category analysis, seller comparisons, and review collection. The bounded modes let you request only the surfaces you need, while the market setting keeps Walmart US and Walmart Canada results separate.

What data can it extract?

The dataset uses one row per source-backed product, detail, category, best-seller, deal, or review record.

Field	Description
row_type, mode, market, query, position	Row classification, source mode, market, query, and result position.
item_id, us_item_id, name, brand, url	Public product identity and canonical product URL.
price, original_price, currency, rollback	Source-provided pricing fields when available.
in_stock, availability_status	Source-provided availability fields.
seller_name, rating, review_count	Public seller and review summary values.
image_url, images, category, badges	Public product media and category metadata.
model_number, upc, specifications, breadcrumbs	Additional product-detail fields when exposed.
review_id, author, title, text, date, verified_purchase, helpful_count	Review fields when review mode is selected.
source_website, source_url, source_transport, extraction_method, source_blocked	Source provenance and access status.

The dataset contains only source-backed product, category, detail, best-seller, deal, and review records. Pages with no usable source data and permanent request failures are reported in run logs and counted in OUTPUT; they do not create diagnostic dataset rows. Missing source values remain absent or null.

How to scrape Walmart

1. Select one or more values in dataTypes.
2. Supply searchQueries for search mode, categoryUrls for category mode, productIds or productUrls for product/review modes, or bestsellersDealsCategory for best-seller/deal modes.
3. Choose market: "US" or market: "CA" and an optional sort order.
4. Set bounded result, page, review, and concurrency limits.
5. Start the Actor, inspect the run logs and OUTPUT summary, and export the Dataset.

Input example

{
  "dataTypes": ["search"],
  "searchQueries": ["wireless keyboard"],
  "market": "CA",
  "sort": "best_match",
  "maxResultsPerType": 40,
  "maxPages": 5,
  "maxConcurrency": 1,
  "includeProductDetails": false
}

dataTypes supports search, category, product, bestsellers, deals, and reviews. Search mode defaults to air fryer when no query is supplied. URL aliases (startUrls, urls, and searchUrls) and legacy aliases such as query, sortBy, and maxItems remain supported and are validated as official HTTPS Walmart URLs or bounded values.

For category mode, provide one or more official /browse/ or /cp/ URLs. For product and review modes, provide numeric item IDs or official /ip/ URLs. For best-seller and deal modes, provide a category keyword or official category URL.

Cost and limits

The Actor uses normal Apify compute, browser, proxy, and network resources. Each selected mode is bounded by maxResultsPerType (1–500), maxPages (1–25), maxConcurrency (1–5), and maxReviewsPerProduct (1–1,000). Search queries are limited to 20 unique values, and URL arrays have explicit upper bounds. Direct browser access is the default; the optional Apify Proxy configuration is available when authorized and required.

Walmart pages can vary by market and can return access challenges. The Actor reports these outcomes in logs and OUTPUT instead of using a CAPTCHA solver, third-party reader, login, stealth behavior, or challenge interaction. If some source-backed rows were stored, later request or optional summary errors keep the Apify run SUCCEEDED and set dataStatus: "PARTIAL". A genuine no-results run is SUCCEEDED with dataStatus: "NO_RESULTS"; a run with permanent request failures and no rows is FAILED with dataStatus: "FAILED".

Output example

{
  "source": "w3crawler/walmart-search-scraper",
  "source_website": "https://www.walmart.ca/",
  "source_url": "https://www.walmart.ca/en/search?q=wireless+keyboard",
  "query": "wireless keyboard",
  "mode": "search",
  "market": "CA",
  "source_transport": "playwright_direct",
  "extraction_method": "embedded_search_payload",
  "source_blocked": false,
  "scraped_at": "2026-09-08T12:00:00.000Z",
  "row_type": "search_result",
  "position": 1,
  "us_item_id": "6D216IVRTJAC",
  "item_id": "6D216IVRTJAC",
  "name": "Public Walmart product",
  "brand": "Example Brand",
  "price": 84.93,
  "original_price": 99.99,
  "currency": "CAD",
  "rollback": false,
  "rating": 4.5,
  "review_count": 12,
  "in_stock": true,
  "seller_name": "Example Seller",
  "availability_status": "Available",
  "image_url": "https://i5.walmartimages.ca/asr/example.jpeg",
  "images": ["https://i5.walmartimages.ca/asr/example.jpeg"],
  "url": "https://www.walmart.ca/en/ip/Public-Product/6D216IVRTJAC"
}

Product-detail rows may also include model, UPC/GTIN, specifications, breadcrumbs, and bounded source data. Review rows use row_type: "review" and include review fields when Walmart exposes them. The OUTPUT key-value record contains per-mode counts, field counts, pages visited, source-issue counts, permanent failure count, and run status.

You can download the dataset in various formats such as JSON, HTML, CSV, or Excel.

Tips for better results

• Set market explicitly when collecting Canadian results so prices and source URLs are interpreted correctly.
• Keep maxResultsPerType small while testing a new mode.
• Use canonical Walmart URLs and inspect the source page when relying on price or availability.
• Review source_transport, extraction_method, and source_blocked when comparing runs.
• Treat absent fields as source limitations rather than inferred values.
• Use proxy settings only when authorized and permitted for the target market.

FAQ

Why did a request produce no dataset row?

The source page may have returned an access challenge, an empty result, malformed public data, or a non-supported page. Inspect the run logs and the OUTPUT status, dataStatus, diagnosticCount, and failureCount. Empty results without permanent request failures finish as SUCCEEDED with dataStatus: "NO_RESULTS"; permanent failures without rows fail the run; runs with usable rows remain SUCCEEDED and use dataStatus: "PARTIAL" when issues occurred.

Can I collect Walmart Canada data?

Yes. Set market to CA and use official Walmart.ca category or product URLs where a URL input is required. Currency and availability are preserved from the source.

Does the Actor use a Walmart API or private account?

No. It reads public page content through bounded browser requests and does not log in or access private systems.

Support, legal notice, and non-affiliation

For support, inspect the run log, OUTPUT, and the Issues tab before reporting a reproducible problem. Use the API tab for programmatic access and the export controls for downstream analysis.

The Actor collects public page data only. Use it for a legitimate purpose, follow Walmart’s terms and applicable law, respect privacy obligations, and honor technical access restrictions. Public product and review information may still have legal or policy constraints in your jurisdiction.

This independent community tool is not sponsored, endorsed, or administered by Walmart Inc. or its affiliates. Walmart and related marks belong to their respective owners.

Local verification

From this Actor directory, run npm test, npm run lint, npm run schema, and apify run --purge. Then run npm run validate:dataset against generated local storage. The validator checks only source-backed dataset rows and reconciles their count with OUTPUT; successful empty results are allowed, while a genuine no-row request failure fails the run.