Steam Reviews & Game Data Scraper

Get Steam user reviews and game data in one run, straight from Steam's official public endpoints: no login, no API key, no proxy. List games by app ID, store URL or name, choose your filters, and get clean rows ready for spreadsheets, sentiment analysis, dashboards or AI tools.

• Full review history (newest first or recently updated), with filters for language, positive/negative, purchase type and an exact date range
• Game records (optional): price, discount, release date, developers, publishers, genres, top user tags, Metacritic, Steam review score and counts, players online right now
• Privacy by default: reviewer Steam IDs are replaced with a hashed ID unless you turn raw IDs on
• Hard limits on reviews per game, reviews per run and money. You never pay for games that weren't found, duplicates or reviews outside your filters. A run that saves nothing costs nothing, not even the start fee.

Why this one is better

Steam retired its old review endpoint (store.steampowered.com/appreviews/...?json=1) on 22 October 2026; it now redirects to the Steam home page. This Actor is built on its replacement, IUserReviewsService/GetAppReviews, and sends every filter as the number Steam expects. That detail matters: when filters are sent as words ("recent", "all"), Steam silently ignores them and falls back to its short "helpful" list. With the correct values this Actor walks the whole history: in our test it paged through all 12,722 of 12,722 English reviews of A Short Hike, back to its release in July 2019.

Comparison with the two most-used Steam review Actors on Apify (their public Store pages, plus one small test run of each with the same input on 3 November 2026: Elden Ring, English, negative only, 14-20 Nov 2023):

	This Actor	automation-lab / Steam Game Reviews Scraper	memo23 / Steam Reviews Scraper
Test result (Steam reports 28 matching reviews)	28 reviews	0 reviews; the log shows it still calls the retired appreviews endpoint ("Invalid JSON after 3 retries"). The $0.003 start fee was charged. Last build 2 Sep 2026.	Works: 25 reviews returned in our run (its own log reported 28 matching)
Price per 1,000 reviews	$0.49 (flat)	$0.575 on Free plan, down to $0.14 on Diamond	$1.00
Start fee	$0.0025, only if the run saves at least one row	$0.003 every run	$0.005 per GB of memory, every run
Input	App IDs, store/community URLs and game names	App IDs, URLs and game names	App IDs and URLs
Game data (price, discount, tags, genres, Metacritic, review score)	Yes, as separate game records	Optional game info attached to reviews (no price, tags or players listed)	No
Current player count	Yes	No	No
Date range	Applied by Steam, works with "Newest first" (full history)	README: dates only work with the "most helpful" sort and it switches to it	Yes, exact window
Reviewer Steam IDs	Hashed by default; raw ID, profile and review links only if you turn them on	Raw Steam ID and profile URL in every row	Raw Steam ID and profile URL in every row
Spend limit	"Max spend (USD)" input plus Run options' maximum cost, checked before every row	Run options only	Run limit shared across games
Off-topic (review-bomb) filter	No	No	Yes ("includeOffTopic")

Prices and features can change; check each Actor's page for the latest. Higher Apify plans get lower prices from automation-lab's tiered pricing, so on Silver plans and above it is cheaper per review than this Actor (when it returns data).

How to use it

1. Under Games, add one game per line: an app ID (1245620), a store URL (https://store.steampowered.com/app/1245620/ELDEN_RING/) or a name (Elden Ring).
2. Set Max reviews per game (default 100). Set it to 0 with Include game details on to get only game records.
3. Optional: pick a language, positive/negative only, purchase type, sort order or a date range.
4. Optional: set Max spend (USD) to cap the cost.
5. Press Start. Rows appear in the Output tab while the run is going.

For a weekly report, schedule the Actor with Only the last N days set to 7.

What you get

Rows start with recordType (review or game). Rows of the same type always have the same columns in the same order; empty values are null, numbers are numbers and dates are ISO 8601 in UTC.

Review row (real row from a test run on 2026-11-04, review text shortened):

{
  "recordType": "review",
  "gameName": "ELDEN RING",
  "appId": 1245620,
  "recommended": true,
  "sentiment": "positive",
  "review": "What can I even say that hasn't been said already? Play this game if you haven't already. It's not for everyone, but everyone should at least try it out.",
  "language": "english",
  "createdAt": "2026-11-05T03:59:15.000Z",
  "updatedAt": "2026-11-05T03:59:15.000Z",
  "votesHelpful": 0,
  "votesFunny": 0,
  "helpfulnessScore": 0.5,
  "commentCount": 0,
  "authorPlaytimeAtReviewHours": 546.3,
  "authorPlaytimeForeverHours": 546.3,
  "authorPlaytimeLastTwoWeeksHours": 0,
  "authorDeckPlaytimeAtReviewHours": null,
  "authorNumReviews": 4,
  "authorLastPlayedAt": "2026-06-13T04:22:37.000Z",
  "steamPurchase": false,
  "receivedForFree": false,
  "writtenDuringEarlyAccess": false,
  "primarilySteamDeck": false,
  "refunded": false,
  "edited": false,
  "developerResponse": null,
  "developerRespondedAt": null,
  "hardwareOs": null,
  "hardwareCpu": null,
  "hardwareGpu": null,
  "hardwareRamMb": null,
  "hardwareVramMb": null,
  "recommendationId": "239407584",
  "authorId": "2342fa57770bf7d74bcb",
  "authorSteamId": null,
  "authorProfileUrl": null,
  "reviewUrl": null,
  "gameReviewsUrl": "https://steamcommunity.com/app/1245620/reviews/",
  "scrapedAt": "2026-11-05T04:12:24.798Z"
}

Game record (recordType: "game", when Include game details is on; real row, long lists shortened):

{
  "recordType": "game",
  "gameName": "ELDEN RING",
  "appId": 1245620,
  "appType": "game",
  "price": 59.99,
  "priceBeforeDiscount": 59.99,
  "discountPercent": 0,
  "currency": "USD",
  "priceFormatted": "$59.99",
  "priceCountry": "US",
  "isFree": false,
  "releaseDate": "Feb 24, 2022",
  "releaseDateIso": "2022-02-24",
  "comingSoon": false,
  "reviewScoreLabel": "Very Positive",
  "reviewPercentPositive": 93,
  "reviewCount": 835430,
  "allReviewsTotal": 1161134,
  "allReviewsPositive": 1080476,
  "allReviewsNegative": 80658,
  "recommendationsTotal": 835430,
  "currentPlayers": 15769,
  "metacriticScore": 94,
  "metacriticUrl": "https://www.metacritic.com/game/pc/elden-ring?ftag=MCD-06-10aaa1f",
  "developers": [
    "FromSoftware, Inc."
  ],
  "publishers": [
    "FromSoftware, Inc.",
    "Bandai Namco Entertainment"
  ],
  "genres": [
    "Action",
    "RPG"
  ],
  "categories": [
    "Single-player",
    "Multi-player",
    "PvP",
    "Online PvP",
    "Co-op",
    "..."
  ],
  "tags": [
    "Souls-like",
    "Open World",
    "Dark Fantasy",
    "RPG",
    "Difficult",
    "..."
  ],
  "platformWindows": true,
  "platformMac": false,
  "platformLinux": false,
  "requiredAge": 16,
  "supportedLanguages": [
    "English",
    "French",
    "Italian",
    "German",
    "Spanish - Spain",
    "..."
  ],
  "fullAudioLanguages": [
    "English"
  ],
  "dlcCount": 3,
  "achievementsTotal": 42,
  "storeUrl": "https://store.steampowered.com/app/1245620/",
  "website": null,
  "headerImage": "https://shared.akamai.steamstatic.com/store_item_assets/steam/apps/1245620/header.jpg?t=1790290043",
  "shortDescription": "THE CRITICALLY ACCLAIMED FANTASY ACTION RPG. Rise, Tarnished, and be guided by grace to brandish the power of the Eld...",
  "scrapedAt": "2026-11-05T04:12:24.584Z"
}

Views. The Output tab opens in Overview (reviews). There are also Games, Review signals (a compact table without the long text), Reviewer hardware and Detailed (every field) views. Download as JSON, CSV, Excel or HTML, or use the API, Make, Zapier, n8n or the Apify MCP server.

Sorting and filters. Sort picks Steam's order (most recent, recently updated or most helpful). Filters for language, positive/negative, purchase type, dates and Minimum playtime at review are applied before the per-game and total limits; skipped reviews are free. Reviews with the same ID are saved only once.

Changed in v1.1: game rows use gameName (was name) and appType (was type); the nested platforms object is now platformWindows, platformMac, platformLinux; the nested review hardware object is now hardwareOs, hardwareCpu, hardwareGpu, hardwareRamMb, hardwareVramMb; columns are in a fixed order; new minPlaytimeHours filter.

The RUN_SUMMARY record in the key-value store lists, for each game, how many reviews were saved, how many Steam reports for your filters, how many duplicates or out-of-range reviews were skipped (free), which inputs weren't found (free), the total charged and why the run stopped.

Review fields

Field	Meaning
appId, gameName	The game
recommendationId	Steam's review ID (used to remove duplicates)
recommended, sentiment	Thumbs up/down, as true/false and as "positive"/"negative"
review, language	Review text and its language
createdAt, updatedAt, edited	Posted and last edited (UTC)
votesHelpful, votesFunny, helpfulnessScore, commentCount	Community reactions
steamPurchase, receivedForFree, writtenDuringEarlyAccess, primarilySteamDeck, refunded	Steam's flags on the review
authorId	Hashed reviewer ID (see Privacy)
authorNumReviews, authorPlaytime...Hours, authorLastPlayedAt	Reviewer's review count and playtime in hours
developerResponse, developerRespondedAt	The developer's public reply, if any
hardwareOs, hardwareCpu, hardwareGpu, hardwareRamMb, hardwareVramMb	PC specs, only if the reviewer chose to show them
authorSteamId, authorProfileUrl, reviewUrl	Only with Include reviewers' raw Steam IDs on

Game record fields

Field	Meaning
gameName, appType, storeUrl	Name, item type (game, dlc...) and store link
price, priceBeforeDiscount, discountPercent, currency, priceFormatted, priceCountry, isFree	Current price in the store country you chose
releaseDate, releaseDateIso, comingSoon	Release information
developers, publishers, genres, categories, tags	Who made it and how Steam and players classify it
metacriticScore, metacriticUrl	Metacritic score shown on Steam
reviewScoreLabel, reviewPercentPositive, reviewCount	The summary shown on the store page
allReviewsTotal, allReviewsPositive, allReviewsNegative	All reviews in every language, including key activations
currentPlayers	Players in the game at scrape time
platformWindows, platformMac, platformLinux, requiredAge, supportedLanguages, fullAudioLanguages, shortDescription, website, headerImage, dlcCount, achievementsTotal, recommendationsTotal	Other store data

Pricing

Pay per event. You only pay for rows saved to your dataset, plus a small start fee once the run has saved something. No platform usage charges on top.

Event	Price	Per 1,000
Review (one review saved)	$0.00049	$0.49
Game record (one game's details saved)	$0.00077	$0.77
Actor start (once per run, only if at least one row is saved)	$0.0025

Examples: 50 reviews of one game = $0.0025 + 50 × $0.00049 = $0.027. 1,000 reviews = $0.49 + $0.0025. Details for 100 games, no reviews = $0.077 + $0.0025 = $0.08.

Never charged: games or names that weren't found, invalid URLs, duplicate reviews (same review ID), reviews outside your date range, reviews without text, failed requests, and runs that save nothing at all.

Input options

Option	Default	What it does
Games	(required)	App IDs, Steam store/community URLs or game names, one per line
Max reviews per game	100	Stop after this many reviews per game. 0 = no reviews (game records only)
Max reviews in total	0 (off)	Hard limit for the whole run
Include game details	off	Add one game record per game
Review language	All languages	One Steam language, or all
Review type	All	Positive only, negative only, or both
Purchase type	All reviewers	Bought on Steam, not bought on Steam (keys, gifts), or all
Sort order	Newest first	Newest first, recently updated (both full history), or most helpful (short list)
Reviews written on or after / on or before	none	Date range (UTC, whole days)
Only the last N days	0 (off)	Shortcut for scheduled runs
Minimum playtime at review (hours)	0 (off)	Keep only reviews written after at least this much playtime; skipped reviews are free
Store country for prices	us	Two-letter country code for prices and currency
Include reviewers' raw Steam IDs	off	Adds raw Steam ID, profile URL and review URL
Your own hashing secret	none	Your own key for the hashed reviewer IDs
Max spend (USD)	none	Stop before charging more than this

Example input:

{
  "games": ["1245620", "https://store.steampowered.com/app/1055540/A_Short_Hike/", "Hades"],
  "maxReviewsPerGame": 200,
  "language": "english",
  "reviewType": "negative",
  "startDate": "2026-01-01",
  "includeGameDetails": true,
  "maxSpendUsd": 1
}

FAQ

Do I need a Steam account or API key? No. Everything comes from Steam's public endpoints that anyone can open without logging in.

Can it really get all reviews of a big game? Yes, with Newest first or Recently updated sorting. Set Max reviews per game high enough. Big games take a while: Steam returns 100 reviews per request, and the Actor waits at least 0.3 seconds between requests to be polite, so expect roughly 8,000-15,000 reviews per minute (our test run saved 3,000 Elden Ring reviews in 23 seconds). Keep an eye on the cost: 100,000 reviews cost $49.

Why does "Most helpful" return only a few hundred reviews? Steam caps that list itself, so the Actor can't page past it. For complete data, use Newest first.

How does name search work? The Actor uses Steam's store search and picks an exact name match if there is one, otherwise the top result. The run log shows which game was picked and the other close matches. If the wrong game is picked, use the app ID (the number after /app/ in the store URL).

Why are there fewer reviews than the number on the store page? The store page counts reviews in all languages from Steam purchasers. Your filters (language, type, purchase type, dates) narrow that down. The log and RUN_SUMMARY show how many reviews Steam reports for your exact filters.

Why does "Bought on Steam" return nothing for a free game? Steam doesn't count free-to-play installs as purchases. Use All reviewers for free games.

How do I make sure I never spend more than a set amount? Set Max spend (USD), or Maximum cost per run in Run options (the lower one wins). The Actor checks the limit before saving each batch and stops before it would go over. Max reviews per game and Max reviews in total are extra hard stops.

What happens if Steam is slow or limits requests? The Actor retries network errors, timeouts and "too many requests" answers with growing pauses (up to 6 attempts) and follows Steam's Retry-After header. If a game still fails, it keeps what was saved, notes the problem in RUN_SUMMARY and moves on to the next game. Failed requests are never charged.

What is the hashed reviewer ID? authorId is an HMAC-SHA256 hash of the reviewer's Steam ID with a secret key. The same reviewer gets the same authorId in all your runs, so you can count repeat reviewers, but the ID can't be turned back into a Steam profile. By default the key is tied to your Apify account; set Your own hashing secret to control it yourself.

Is it legal? What about privacy? The Actor only reads data that Steam publishes openly without a login, through Steam's own public endpoints, at a polite rate. Reviews are written by real people, so they are personal data in many countries (for example under the GDPR) even though they are public. That's why reviewer Steam IDs are hashed by default and no names or avatars are collected. Review text can still contain personal details, so use the data for a clear purpose (such as product research or sentiment analysis), keep only what you need, and don't use it to contact or profile individual players. Steam's terms and copyright in the review texts also apply to what you do with the data. This Actor is not affiliated with or endorsed by Valve. If in doubt, ask a lawyer.

Will a run ever fail because one game is wrong? No. Invalid app IDs, unknown names and URLs without an app ID are logged, listed in RUN_SUMMARY and skipped for free; the other games are still collected and the run finishes normally.

Something doesn't work. Open an issue on the Issues tab with the run link. Every run ends with a clear status message and a RUN_SUMMARY.