Microsoft Store Scraper — App Details, Search & Reviews (No API Key)

Scrape the Microsoft Store (Windows apps & games) into clean, structured JSON — app details with all-time and 7/30-day rating windows, keyword search, and customer reviews with helpful votes and device info. Any country market. No account, no login, no API key.

✅ Last verified working: 2026-06-11. Monitored by an automated daily canary. When Microsoft changes the Store's endpoints, we patch fast and log it in the changelog below — so your pipeline doesn't silently break.

---

Why this Microsoft Store scraper

Together with our Apple App Store and Google Play scrapers, this completes all three major app stores in one consistent toolset — same modes, same pricing model, same reliability standard. And the Microsoft Store data has fields the other stores don't expose:

• Rating windows — every details record carries the average + count for all-time, last 7 days, and last 30 days. That's a momentum signal: a 4.6 all-time / 3.8-last-30-days app is in trouble; the reverse is a turnaround.
• Reviews with helpful votes + device family — each review includes helpfulPositive/helpfulNegative (community-validated complaints!) plus the reviewer's deviceFamily and osVersion.
• Any country market — market/language inputs localize availability, prices, descriptions, ratings, and review pools (e.g. DE/de-de).
• Batched details — up to 20 apps per request under the hood: fast and cheap.
• Per-item isolation + retries — a bad ID or transient error is skipped and logged, never fatal.
• Stable, versioned output (_schemaVersion) + a daily canary — that "last verified working" date is real.

What data you get

Details rows (one per app):

Field	Type	Description
productId	string	Microsoft Store product ID (12 chars, starts with 9).
title / publisher	string	App name and publisher.
description / shortDescription	string | null	Full and short listing text (localized per language).
category	string | null	Store category (e.g. Social).
rating	object	{average, count} — all-time.
rating7Days / rating30Days	object	{average, count} for the recent windows — the momentum signal.
price	object	{amount, currency, isFree} for the selected market.
imageUrl / imagesCount	string | null, number	Primary logo + total image count.
releaseDate / lastModified	string | null	ISO timestamps.
storeUrl	string	Public listing URL.

Search rows: productId, title, publisher, averageRating, ratingCountText (the Store's humanized count, e.g. "332K") + ratingCountApprox (parsed number), displayPrice, isFree, categories, productType (Apps/Games), hasIAPs, minimumAge, releaseDate, lastModified, storeUrl, _searchTerm.

Review rows: productId, reviewId, reviewerName, rating (1–5), title, text, helpfulPositive, helpfulNegative, market, locale, osVersion, deviceFamily, submittedAt, isRevised.

Use cases

• App competitive intelligence. Track competitors' rating windows (is their last update tanking?), review complaints ranked by helpful votes, and price/IAP changes — across any country market.
• Cross-store ASO research. Combine with our Apple and Google Play scrapers for the full three-store picture of an app brand or category, in one normalized pipeline.
• Review mining. The helpful-vote counts surface the complaints users themselves validated — gold for product teams and reviewers. deviceFamily splits desktop vs Xbox vs mobile feedback.
• Market research. Search any keyword per market and compare which apps win the same query in the US, Germany, or Japan.

Input

Field	Type	Required	Notes
mode	string	✅	details (default) / search / reviews.
productIds	string[]	for details/reviews	12-char Store IDs starting with 9 (e.g. 9NKSQGP7F2NH) or apps.microsoft.com/detail/... URLs. XP... winget IDs are desktop installers, not Store products — they're skipped with a warning.
searchTerms	string[]	for search	Free-text queries; one result set per term.
maxSearchResults	integer	—	Cap per term (1–20). Default 10. The Store's search endpoint returns ~20 results per query.
maxReviewsPerApp	integer	—	Cap per app (1–500), newest first. Default 50. Deep pagination supported.
market	string	—	2-letter country (default US).
language	string	—	Language tag, lowercase (default en-us).

{
  "mode": "details",
  "productIds": ["9NKSQGP7F2NH", "9NCBCSZSJRSB", "9WZDNCRFJ3TJ"]
}

Output sample

A real record from the default run:

{
  "_type": "app",
  "_schemaVersion": "1.0",
  "_source": "microsoft-store",
  "productId": "9NKSQGP7F2NH",
  "title": "WhatsApp",
  "publisher": "WhatsApp Inc.",
  "description": "WhatsApp from Meta is a 100% free messaging app. It's used by over 2B people in more than 180 countries. ...",
  "shortDescription": null,
  "category": "Social",
  "rating": { "average": 4.4, "count": 2294755 },
  "rating7Days": { "average": 4.4, "count": 15549 },
  "rating30Days": { "average": 4.4, "count": 68880 },
  "price": { "amount": 0, "currency": "USD", "isFree": true },
  "imageUrl": "https://store-images.s-microsoft.com/image/apps.39936.13655054093851568.04a4d8fc-dd15-4e4e-bd00-4bbe4247c42e.5fccebf1-7bd7-4999-87f1-863ef313827f",
  "imagesCount": 24,
  "releaseDate": "2017-02-01T01:35:55.265Z",
  "lastModified": "2026-05-27T20:49:24.181Z",
  "storeUrl": "https://apps.microsoft.com/detail/9NKSQGP7F2NH",
  "_scrapedAt": "2026-06-11T12:19:16.266Z"
}

Use with AI agents (MCP)

This scraper is callable as a tool by AI agents via the Apify MCP server — so assistants like Claude, Cursor, and VS Code can run it and use the structured results mid-conversation, with no glue code.

Simplest path: in a recent Claude Desktop, add the hosted server https://mcp.apify.com under Settings → Connectors and authorize with OAuth. For any config-file client:

{
  "mcpServers": {
    "apify": {
      "command": "npx",
      "args": ["-y", "@apify/actors-mcp-server"],
      "env": { "APIFY_TOKEN": "YOUR_APIFY_TOKEN" }
    }
  }
}

Then ask in plain language, e.g. "Pull the Microsoft Store ratings and the 20 most helpful recent reviews for WhatsApp." — the agent calls freshactors/microsoft-store-scraper with the right input and reads back the JSON. Full setup: Apify MCP docs.

Pricing

Pay-per-event — no subscription, no per-run start fee. You pay only for the rows you receive:

Event	Price (USD)	When it fires
App details fetched	$0.0025	Once per app details record returned.
Search result fetched	$0.001	Once per search result row returned.
Review fetched	$0.0001	Once per review returned.

Worked examples:

• Full details on 100 apps: 100 × $0.0025 = $0.25
• 10 search queries × 10 results: 100 × $0.001 = $0.10
• 500 reviews of one app: 500 × $0.0001 = $0.05

(Apify platform usage is billed separately per their standard rates.)

FAQ

Is scraping the Microsoft Store legal? This actor reads public, non-personal listing data from the Store's own public endpoints — the same data the Store app shows any user. Review records contain only the public display name a reviewer chose to publish. Use responsibly and in line with applicable terms and laws.

Do I need an API key or login? No. No Microsoft account, token, or key is involved.

How do I find an app's product ID? Open the app on apps.microsoft.com — the ID is the last path segment: apps.microsoft.com/detail/9NKSQGP7F2NH. You can paste the whole URL too.

Why was my XP... ID skipped? IDs starting with XP come from the winget package index — they're classic desktop installers, not Store catalog products, and the Store's catalog has no data for them. Only 9... Store product IDs resolve.

What are the rating windows? The Store tracks the rating average and count for all-time, the last 7 days, and the last 30 days. Comparing them shows momentum — e.g. an app whose last-30-days average is far below all-time is regressing.

Can I get reviews in other languages/markets? Yes — set market + language (e.g. DE + de-de); the review pool, descriptions, and prices follow the market.

Why is the search result count capped at 20? The Store's search endpoint returns about 20 results per query with no pagination cursor — so we cap honestly rather than promise more. Use more specific terms to surface different slices.

Is ratingCountApprox exact? No — search results carry the Store's humanized count ("332K"); we parse it to an approximate number and keep the original text. For exact counts, run details mode (e.g. WhatsApp: 2294755).

How often is the data updated? It's live — every run queries the Store's current data. The actor itself is monitored by a daily canary so parsing stays correct as the endpoints evolve.

Other FreshActors tools

FreshActors actor	What it scrapes
App Store Scraper	Apple App Store — app details, search, reviews
Google Play Scraper	Google Play — app details + reviews
VS Code Marketplace Scraper	VS Code extensions — install/trending stats, search, reviews
Shopify App Store Scraper	Shopify App Store — app details, reviews, discovery
Redfin Scraper	Redfin — US real-estate listings
Greenhouse & Lever Jobs Scraper	Greenhouse + Lever — normalized job postings
Workable Jobs Scraper	Workable — normalized job postings, full descriptions
SmartRecruiters Jobs Scraper	SmartRecruiters — normalized job postings, full descriptions
Recruitee Jobs Scraper	Recruitee — normalized job postings, full descriptions + salary
Teamtailor Jobs Scraper	Teamtailor — normalized job postings, full descriptions
Personio Jobs Scraper	Personio — normalized job postings, departments + seniority

Reliability

A scheduled daily canary fetches known apps, validates the normalized schema (rating windows numeric, search IDs valid, review pagination advancing, helpful votes present), and alerts the moment a parse changes. The Store's endpoints serve Microsoft's own Store client and can change without notice — when they do, we patch fast and record it below. Every record carries _schemaVersion and _scrapedAt.

Changelog

• 2026-06-11 — v1.0 launch. Details (batched, rating windows, any market) / search (full consumer coverage) / reviews (helpful votes, device family, deep offset pagination); 9* ID validation with clear winget-ID warnings; per-item isolation; daily canary monitoring.

Legal note: Reads public, non-personal listing data from the Store's public endpoints (the same data the Store client serves any user). No login, no private data. Use responsibly and in line with applicable terms and laws.

Found a problem? Open an issue on the actor — issues are answered fast.