Steam Reviews Scraper - Official API, No Key

Export Steam user reviews for any game via Valve's official IUserReviewsService API (the replacement for the retired appreviews endpoint) plus store app details. Filter by language, sentiment, purchase type, playtime and date. No login, no proxy, no browser.

What does this Actor do?

Steam is the biggest PC game store in the world, and its user reviews are the most honest signal of how players feel about a game. On 22 October 2026 Valve switched off the old store.steampowered.com/appreviews?json=1 endpoint that most scripts and scrapers relied on - it now silently redirects to the store homepage. This Actor already uses the official replacement, IUserReviewsService/GetAppReviews, which returns public reviews without any API key, and enriches each review with the game's store details (name, developer, publisher, price, genres).

Key features:

• Plain HTTP against official public endpoints - no browser, no proxy, no login, runs on 128 MB memory.
• Pay only for results: one charge per row saved; your maximum charge per run is always respected.
• maxItems hard cap (default 100) and bounded pagination - runs can never loop forever.
• Incremental mode via a since-date so scheduled runs only fetch new data.
• Polite rate limiting, retries with exponential backoff and Retry-After support.
• Clean, flat, consistently named JSON fields ready for Excel, Google Sheets, BigQuery or your API.

Who is it for?

• Game studios and publishers - track sentiment after a patch, sale or launch and catch negative review bombs early.
• Market researchers - compare review volume and positivity across competitors in a genre.
• Data scientists / NLP - build labelled sentiment datasets (each review carries votedUp) in dozens of languages.
• Steam Deck & hardware analysis - every review has primarilySteamDeck and, when Steam exposes it, the reviewer's OS and GPU.
• Community managers - export the most helpful reviews to plan FAQ, roadmap and store-page copy.

How it works

For each app ID the Actor calls appdetails once (optional) and then pages through GetAppReviews 100 reviews at a time using Steam's cursor. Requests are spaced by delayMs, retried with exponential backoff on 429/5xx, and duplicates are dropped. When you set sinceDate, the date range is sent to Valve's server so only new reviews are downloaded - ideal for a daily or hourly schedule.

Input

Field	Type	Default	Description
appIds	array	-	Numeric app IDs (e.g. 730) or store URLs like https://store.steampowered.com/app/1145360/Hades/.
maxItems	integer	100	Hard cap on reviews saved and charged across all apps.
maxItemsPerApp	integer	0	0 = no per-app cap (only the total cap applies).
filter	string	"recent"	recent = newest first (best for incremental runs), updated, helpful, funny.
language	string	"all"	Steam API language code (english, schinese, german...) or 'all'.
reviewType	string	"all"	all, positive or negative.
purchaseType	string	"all"	all, steam (bought on Steam) or non_steam_purchase (keys).
sinceDate	string	-	ISO date (YYYY-MM-DD). Uses the API's server-side date range so incremental runs stay cheap.
minPlaytimeHours	integer	0	Skip reviews from players with less playtime.
includeAppDetails	boolean	true	Adds name, developer, publisher, price, genres from the store appdetails API (1 extra request per app).
delayMs	integer	500	Polite rate limit.

Example input

{
  "appIds": [
    "1145360",
    "https://store.steampowered.com/app/730/"
  ],
  "maxItems": 500,
  "filter": "recent",
  "language": "english",
  "sinceDate": "2026-10-01"
}

Output

Each result is one row in the default dataset. You can download it as JSON, CSV, Excel, XML or HTML, or read it through the Apify API. A real row from a test run on 26 October 2026:

{
  "appId": 730,
  "appName": "Counter-Strike 2",
  "appType": "game",
  "developers": [
    "Valve"
  ],
  "publishers": [
    "Valve"
  ],
  "releaseDate": "Aug 21, 2012",
  "priceUsd": 0,
  "genres": [
    "Action",
    "Free To Play"
  ],
  "reviewId": "238567914",
  "language": "russian",
  "review": "Ебаное говно не регает",
  "votedUp": true,
  "votesUp": 0,
  "votesFunny": 0,
  "weightedVoteScore": 0.5,
  "commentCount": 0,
  "steamPurchase": true,
  "receivedForFree": false,
  "writtenDuringEarlyAccess": false,
  "primarilySteamDeck": false,
  "refunded": false,
  "hardwareOs": null,
  "hardwareGpu": null,
  "createdAt": "2026-10-25T19:55:28.000Z",
  "updatedAt": "2026-10-25T19:55:28.000Z",
  "authorSteamId": "76561199813609579",
  "authorNumReviews": 1,
  "authorPlaytimeHours": 413.4,
  "authorPlaytimeAtReviewHours": 413.4,
  "appReviewScoreDesc": "Very Positive",
  "appTotalReviews": 9944731,
  "url": "https://steamcommunity.com/profiles/76561199813609579/recommended/730/"
}

Tips

• Use filter: recent plus sinceDate on a schedule for cheap incremental monitoring.
• Use filter: helpful to get the reviews Steam shows first on the store page.
• language: all returns every language; use codes like schinese, german, brazilian to focus.
• Playtime fields are converted to hours to make filtering easy in spreadsheets.

Pricing

This Actor uses pay-per-event pricing: about $0.50 per 1,000 results. There is no monthly rental fee. Platform usage is minimal because the Actor runs on 128 MB with plain HTTP requests. Set Maximum cost per run in the run options and the Actor stops cleanly as soon as the limit is reached - you are never charged beyond it.

Scheduling and integrations

Create an Apify Schedule (for example daily at 08:00) and pass a since-date to get only new items each time. Connect the dataset to Google Sheets, Slack, Zapier, Make, n8n or a webhook to receive results automatically, or call the Actor from Python or JavaScript using the Apify API client.

Limitations

Only public reviews are available. Anonymous calls are rate-limited by Valve and responses may be cached up to 10 minutes, so very fresh reviews can appear with a short delay. appTotalReviews reflects the active filters (e.g. date range).

FAQ

Do I need an account or API key on the source site? No. Only public, unauthenticated endpoints are used.

Do I need proxies? No. The Actor makes a small number of well-spaced requests to official APIs, so datacenter IPs work fine.

Why did my run return fewer items than maxItems? The source simply had no more items matching your filters and date range, or your maximum cost per run was reached.

Is this legal? The Actor only reads publicly available data through documented public endpoints. You are responsible for using the data in line with the source's terms and applicable laws such as GDPR, especially for personal data.

Something broke? Open an issue on the Actor's Issues tab with your run ID and we will fix it quickly.