Steam Reviews & Game Details Scraper (official store API)

Scrape Steam user reviews for any game, plus game details, using Steam's public store endpoints (store.steampowered.com/appreviews and store.steampowered.com/api/appdetails). No login, no cookies, no browser, no Steam Web API key. One dataset row per review, plus one optional free row per game.

Great for: game market research, player-sentiment tracking, competitor analysis, patch and launch monitoring, building review datasets for NLP, and finding what players praise or complain about.

How it works

1. You give app IDs (730), store links or game names (Stardew Valley: the exact store title wins, otherwise the top store-search hit, so check appName). An ID, a URL and a name for the same game count once.
2. If includeAppDetails is on, the actor first fetches the game's details from the store API (US store, English). An app that doesn't exist gets an error row.
3. It then pages through Steam's review endpoint (up to 100 reviews per request) with your sort, language, review-type and purchase-type filters until maxReviewsPerApp is reached or Steam has no more reviews.
4. Each review is normalized into one row and deduplicated by review ID. The game-details row, with the review score summary, is written after the reviews.

Up to 2 games are processed in parallel, with a small delay between requests. HTTP 429 and 5xx responses are retried with backoff (up to 4 retries).

Features

• Paste app IDs (730) or store links (https://store.steampowered.com/app/730/...).
• Sort by most recent, recently updated or most helpful (with optional day range).
• Filter by language, positive/negative, and Steam purchase vs. key activation.
• Per review: thumbs up/down, full text, language, created/updated dates, helpful & funny votes, weighted score, comment count, Steam purchase / received for free / early access / Steam Deck flags, playtime at review and total playtime (hours).
• Optional free game details row per game: developers, publishers, release date, price (USD), genres, categories, platforms, Metacritic, and the review score summary (e.g. "Very Positive", total positive/negative).
• A bad app ID never crashes the run: you get a row with type: "error" (not charged).

Use cases

Game market research

Pull the newest reviews across a list of competitor games and compare what players complain about, with playtime showing how far in the reviewer was.

Sentiment and launch monitoring

Schedule a run with the recent sort after a patch or sale and watch the positive/negative split and recent review text.

NLP and LLM datasets

Export review text with language, votes and playtime as JSON or CSV. Use language to restrict to one language.

Publisher and developer feedback loops

Filter to negative reviews only, sorted by most helpful, to see the most-endorsed criticisms.

Input

Field	Description
apps	Steam app IDs, store URLs or game names
maxReviewsPerApp	Max reviews per game (default 200, up to 100,000; 0 = details only)
filter	recent (default), updated, or all (most helpful)
language	Steam language code: all (default), english, german, schinese, ...
reviewType	all (default), positive, negative
purchaseType	all (default), steam, non_steam_purchase
dayRange	Last N days (max 365); only with filter: all
includeAppDetails	Add one free game-details row per game (default true)

Output

{
  "type": "review", "appId": 730, "appName": "Counter-Strike 2", "recommendationId": "236483094",
  "url": "https://store.steampowered.com/app/730/#app_reviews_hash",
  "votedUp": false, "text": "...", "language": "russian",
  "createdAt": "2026-09-29T16:50:29.000Z", "updatedAt": "2026-09-29T16:50:29.000Z",
  "votesUp": 0, "votesFunny": 0, "weightedVoteScore": 0.5, "commentCount": 0,
  "steamPurchase": true, "receivedForFree": false, "writtenDuringEarlyAccess": false, "primarilySteamDeck": false,
  "playtimeForeverHours": 42.8, "playtimeAtReviewHours": 42.8, "authorNumReviews": 1,
  "scrapedAt": "...", "error": null
}

Game rows have type: "app" with name, developers, publishers, releaseDate, price, genres, categories, platforms, metacritic and totalReviews (reviewScoreDesc, totalPositive, totalNegative, totalReviews). appName on review rows is filled only when includeAppDetails is on. The review url links to the game's store reviews section, not to the individual review. Privacy: no Steam IDs, usernames, avatars or profile links are returned, only the review content and anonymous playtime stats.

Pricing

Pay per event: $0.0005 per review ($0.50 per 1,000). Apify platform usage is billed separately as usual.

• 200 reviews (the default) for one game = $0.10
• 10 games x 500 reviews = 5,000 reviews = $2.50
• 50,000 reviews = $25.00

Charged: each review row saved. Not charged: game-detail rows, error rows and duplicate reviews. Details-only runs (maxReviewsPerApp: 0) cost no review events. If you set a maximum charge for the run, the actor stops before exceeding it.

Limitations

• dayRange only applies to the "most helpful" sort (a Steam restriction).
• With includeAppDetails: false, an unknown app ID returns zero reviews instead of an error row.
• Counts and votes are a snapshot at scrape time; deleted or hidden reviews are not returned.
• Prices are USD (US store); no regional prices.
• Not affiliated with Valve or Steam. Built and maintained with AI assistance; problems go to the Issues tab.

FAQ

Do I need a Steam account or API key? No. Everything comes from Steam's public store endpoints.

How many reviews can I get? As many as Steam's review pages return for your filters, up to 100,000 per game per run; set maxReviewsPerApp. Very popular games have millions of reviews, so start small.

Can I get only negative reviews, or only one language? Yes, use reviewType and language.

Can I get only game details, without reviews? Yes. Set maxReviewsPerApp to 0; you get the free details row and score summary.

Why are the totals different from the store page? totalReviews follows your language / review type / purchase type filters, and the store page may count reviews differently (e.g. off-Steam activations).

Can I get regional prices? Not in this version; prices are the US store price.

Is this legal / allowed? It reads only public review data through Steam's public store endpoints at modest request rates and returns no personal identifiers. You are responsible for how you use the data.

Related tools

Other actors by the same developer (same flat pay-per-result pricing, no subscription):

• App Store Reviews Scraper – Apple iOS, Multi-Country
• Article Scraper & Text Extractor – Clean Markdown for LLM/RAG
• Job Postings & Career Page Scraper – Workday, Greenhouse, Lever
• Bluesky Posts Scraper – Profiles & Threads, No Login
• Company Enrichment from Domain (Logo, Socials, Tech)
• DNS Lookup & WHOIS Domain Checker — SPF/DMARC, SSL Expiry
• Email Validator & MX Record Checker – Bulk, Disposable, Role
• Bulk PageSpeed Insights, Lighthouse & Core Web Vitals
• PDF Extractor & Parser – Bulk PDF to Text with Metadata
• SEO Audit Crawler & Broken Link Checker – Website Health
• Sitemap Extractor & Bulk URL Status Checker
• Website Technology Detector – Tech Stack, CMS & Analytics
• Website Screenshot API – Bulk Full Page PNG, JPEG & PDF

Integrations

Run it from the Apify API, a schedule, or no-code tools: the Apify apps for Zapier, Make and n8n can start any public actor ("Run Actor") and read its dataset. AI agents can call it through the Apify MCP server.

Ready-made examples

One-click tasks you can run or copy and adapt:

• Latest negative Steam reviews for Counter-Strike 2
• Compare Steam reviews of hit indie games (Hades, Stardew)
• Baldur's Gate 3 Steam reviews in German
• Cyberpunk 2077: recently updated Steam reviews after patches