Steam Reviews & Game Details Scraper — No login

Steam Reviews & Game Details exports Steam game reviews by app ID or store link. Each review comes with the verdict, text, language, playtime, helpful votes and date, and each game gets one row with price, genres, release date and review score.

Use it to study what players like or hate, to watch new reviews after a patch, or to compare games. The smallest run is one app ID, for example 730, and 20 reviews. No Steam login or API key is needed. That prefill run (20 reviews and one game row) costs about $0.01.

Table of contents

• What it does
• How it works
• Quick start
• Use cases
• Input
• Output
• Pricing
• Limits
• Scheduled runs: only new reviews
• API, export and AI agents
• Troubleshooting
• FAQ
• Related Actors
• Support

What it does

You give the Actor Steam app IDs or store links. For each game it reads the public review list of the Steam store and saves one dataset row per review. It can also save one game details row with the current store price, discount, genres, release date, developers and the whole-game review score.

You choose the review language, verdict (Recommended or Not recommended), order (newest, recently updated or most helpful), purchase type and a date period. The Actor stops at your number of reviews per game, at the end of the list, or at your maximum charge.

Good fit

• Read player feedback for your game or a competitor's game in one table.
• Collect only reviews written after a date, for example after a patch or sale.
• Compare price, discount, genres and review score of many games at once.
• Build a sentiment or topic dataset in one language.

Not a fit

• Only the Steam store. Bundles, packages, DLC lists, workshop items and player profiles are not supported.
• Player data stays minimal: you get the reviewer's steamId and the public review link, not names, avatars or friend lists.

👉 Try it now: open the input form

How it works

1. The Actor reads each game's public review list from Steam's official review service, page by page, with your language, verdict, order, purchase type and date period applied by Steam.
2. It stops a game at your review count, at the end of the list or at your start date, and asks the Steam store once per game for price, genres, release date and review score.
3. It saves one row per review and one game row per game to the dataset, the game rows also as the GAMES record, and a RUN_SUMMARY record with the run status and a checkpoint for the next scheduled run.

👉 Try it now: open the input form

Quick start

1. Open the Input tab. Steam app IDs already holds 730, Counter-Strike 2 (box 1 in the picture). The ID is the number in the store link: store.steampowered.com/app/730/.

2. Keep it, or paste store links of your games into Steam store URLs (box 2).

3. Keep Max reviews per game at the prefilled 20 (box 3), then click Start at the bottom of the form.

4. Open the Reviews view of the output and look at the verdict, review text and hours played (box 4 in the picture below). The All rows view adds the game row with price and review score.

   

Expected result: 20 review rows and 1 game row in a few seconds.

👉 Try it now: open the input form

Use cases

Buyer job	Input path	Output entity	Key fields	Next integration
Find top complaints about a game	appIds, reviewType: negative, language: english	review rows	text, votesUp, playtimeHours	Topic or sentiment analysis in Sheets or an LLM
Check reaction to a patch	reviewsSince: "2026-10-01", reviewsUntil: "2026-10-07"	review rows	recommended, createdAt	Share of positive reviews per day
Compare a genre	20–100 appIds, maxReviews: 0	game rows	price, discountPercent, genres, positivePercent	Price and score table
Track a competitor's sale	appIds, countryCode, scheduled	game rows	price, initialPrice, discountPercent	Alert when the discount changes

👉 Try it now: open the input form

Input

Input fields

Field	Plain name	Type	Required	Default / prefill	Allowed values	Effect	Example	Recommendation
appIds	Steam app IDs	array of strings	one of appIds/gameUrls	none (prefill ["730"])	numbers 1–2147483647	Games to read	["730", "570"]	Easiest way to start
gameUrls	Steam store URLs	array of URLs	one of appIds/gameUrls	none	store.steampowered.com/app/…, steamcommunity.com/app/…	Same as app IDs, from links	["https://store.steampowered.com/app/413150/"]	Paste links when you do not know the ID
maxReviews	Max reviews per game	integer	no	100 (prefill 20)	0–50000	Stop after this many reviews for each game. 0 = only the game row	200	Start small, then raise
language	Review language	string	no	all	Steam names (english, russian, schinese, brazilian, koreana …) or codes (en, ru, pt-br)	Only reviews written in this language	english	Set it for text analysis
reviewType	Verdict	enum	no	all	all, positive, negative	Only Recommended or only Not recommended	negative	negative to find problems fast
sortBy	Order	enum	no	recent	recent, updated, helpful	Which reviews come first	helpful	recent for monitoring, helpful for the best-known opinions
reviewsSince	Only reviews since	string	no	none	date 2026-01-31, ISO time, or period 7 days / 30d	Skip older reviews	2026-10-01	Use with scheduled runs
reviewsUntil	Only reviews until	string	no	none	date 2026-03-31 (to the end of that day UTC), ISO time, or period 30 days	Skip newer reviews	2026-10-31	With reviewsSince for a closed period
purchaseType	Purchase type	enum	no	all	all, steam, non_steam_purchase	Only Steam buyers or only keys bought elsewhere	steam	Keep all unless you study key resellers
includeGameDetails	Add game details row	boolean	no	true	true, false	One extra row per game with price, genres and score	false	Turn off if you need reviews only
countryCode	Store country	string	no	us	two letters, e.g. us, de, gb	Price and currency of the game row	de	Games not sold there get no game row

Add at least one app ID or game URL. Up to 100 games per run; duplicates (the same game as an ID and as a link) are processed once. An entry that is not a Steam app ID or store link (for example a bundle link) is skipped with a free error row (errorCode: INVALID_INPUT), and the other games still run; the run then ends as PARTIAL with reason SOME_INPUTS_INVALID. Only when no entry is valid does the run stop with INVALID_INPUT, and you pay only the $0.001 start fee.

Interacting settings

When	Input A	Input B	Effective behavior	Example	Cost/result effect
Monitoring	sortBy: recent	reviewsSince	Stops as soon as reviews get older than the date	"reviewsSince": "7 days"	Fast and cheap: one review page per game when nothing is new
Best reviews of a period	sortBy: helpful	reviewsSince	Steam's most helpful reviews written in that period	"reviewsSince": "2026-01-01"	Charged per saved review only
Closed period	reviewsSince	reviewsUntil	Only reviews created between the two times; Steam applies both	"reviewsSince": "2026-10-01", "reviewsUntil": "2026-10-07"	Fewer rows, so cheaper
Only the game row	maxReviews: 0	includeGameDetails: true	No reviews, one game row per game	"maxReviews": 0	$0.0012 per game
Filters	language, reviewType, purchaseType	maxReviews	Steam applies the filters; maxReviews counts only matching reviews	"language": "english"	Fewer rows when few reviews match

Input examples

Recent negative English reviews of two games

{
  "appIds": ["1091500"],
  "gameUrls": ["https://store.steampowered.com/app/413150/Stardew_Valley/"],
  "language": "english",
  "reviewType": "negative",
  "maxReviews": 200
}

Expected behavior: reads the newest Not recommended English reviews of both games. Expected output: up to 200 review rows per game (recommended: false, text, playtimeHours) plus 2 game rows. Limit or cost note: up to $0.16 for 400 reviews on the Free plan, plus start, game and game-row fees.

Only game details for a list of games

{
  "appIds": ["730", "570", "1245620", "1086940"],
  "maxReviews": 0,
  "countryCode": "de"
}

Expected behavior: no review pages, one store request per game. Expected output: 4 game rows with euro price, discountPercent and reviewScoreDesc. Limit or cost note: $0.0058 in total.

👉 Try it now: open the input form

Output

All rows go to one dataset with two views: Reviews and All rows. The game rows are also saved alone as the GAMES record in the key-value store, so you get a clean game table. Every row has type, source, url and fetchedAt.

Row type	One row per	Main fields
review	review	recommended, text, language, playtimeHours, votesUp, createdAt
game	game (when includeGameDetails is on)	price, discountPercent, genres, releaseDateIso, reviewScoreDesc, positivePercent
not_found	app ID that Steam does not know	appId, message
error	game that failed after retries, or an entry that is not a Steam app ID	appId, input, errorCode, message

Output fields

Review rows (type: "review")

Field	Meaning	Type	Can be empty?	Example / enum	Source / notes
appId	Steam app ID	string	no	"730"	from your input
gameName	Game name	string	yes, if Steam gives no name	"Counter-Strike 2"	store details or review totals
reviewId	Steam review ID, stable key	string	no	"219989924"	Steam recommendationid
steamId	Reviewer's public Steam ID (64-bit)	string	yes	"76561198351194597"
recommended	Recommended (true) or Not recommended (false)	boolean	no	false
text	Review text as written	string	no (may be "")	"The gameplay is nice…"	Steam formatting tags kept
language	Steam language name	string	yes	"english"
playtimeHours	Hours played when the review was written	number	yes	3097.1	Steam minutes ÷ 60, one decimal
playtimeForeverHours	Total hours played now	number	yes	3108.1
playtimeLastTwoWeeksHours	Hours in the last two weeks	number	yes	0
votesUp	"Helpful" votes	integer	yes	940	changes over time
votesFunny	"Funny" votes	integer	yes	57
helpfulScore	Steam helpfulness score 0–1	number	yes	0.91
commentCount	Comments under the review	integer	yes	22
steamPurchase	Bought on Steam	boolean	yes	true
receivedForFree	Got the game for free	boolean	yes	false
refunded	Refunded	boolean	yes	false
writtenDuringEarlyAccess	Written in Early Access	boolean	yes	false
primarilySteamDeck	Played mostly on Steam Deck	boolean	yes	false
developerResponse	Developer's reply	string	yes, no reply	null
developerRespondedAt	Reply time (ISO)	string	yes	null
createdAt	Review time (ISO, UTC)	string	yes	"2026-03-06T15:36:44.000Z"
updatedAt	Last edit time (ISO, UTC)	string	yes	"2026-03-06T15:36:44.000Z"
url	Public review page	string	yes	"https://steamcommunity.com/profiles/…/recommended/730/"
storeUrl	Game store page	string	no	"https://store.steampowered.com/app/730/"
source, fetchedAt	Data source and fetch time	string	no	"steam"

Game rows (type: "game")

Field	Meaning	Type	Can be empty?	Example / enum	Source / notes
appId, gameName, appType	ID, name, Steam type	string	gameName/appType yes	"730", "game"
isFree	Free to play	boolean	yes	true
price, initialPrice	Current and full price in major units	number	yes, if not sold	17.99, 59.99	store of countryCode
currency	ISO currency	string	yes (free games)	"USD"
discountPercent	Current discount	integer	yes	70
priceFormatted	Price as shown in the store	string	yes	"$17.99"
genres, categories	Steam genres and features	array	no (may be empty)	["RPG"]
releaseDate	Release date as Steam shows it	string	yes	"Dec 9, 2020"
releaseDateIso	Same as YYYY-MM-DD when Steam gives a full date	string	yes	"2020-12-09"
comingSoon	Not released yet	boolean	yes	false
developers, publishers, platforms	Who made it, where it runs	array	no (may be empty)	["windows"]
metacriticScore	Metacritic score	integer	yes	86
recommendationsTotal	Steam recommendations counter	integer	yes	5315663
reviewScore	Steam score 1–9	integer	yes	8
reviewScoreDesc	Score label	string	yes	"Very Positive"
totalPositive, totalNegative, totalReviews	Whole-game review counts	integer	yes	9927956
positivePercent	Positive share, %	number	yes	85.8
requiredAge	Age limit	integer	yes	0
shortDescription, headerImage, website	Store text, image and site	string	yes
countryCode	Store country used for price	string	no	"us"
url, source, fetchedAt	Store page, source, fetch time	string	no

Issue rows (type: "not_found" or "error")

appId, gameName, input (what you entered), errorCode, message, url, source, fetchedAt. On an INVALID_INPUT row appId and url are null, because the entry was not a Steam app ID. Issue rows are not charged; each app ID the run starts costs only the $0.0002 game-processed fee.

Output entities and surfaces

Situation	Dataset/API/KVS/live view	Result shape	User action
Normal item	Dataset views Reviews and All rows, GAMES record	review and game rows	Open, export or read reviewId
Domain-empty result	Dataset and RUN_SUMMARY	no review rows, stopReason: end_of_reviews or since_reached	Widen filters or the date
Unknown app ID	Dataset, RUN_SUMMARY	not_found row, status NOT_FOUND when every ID is unknown	Fix the ID
Partial result	RUN_SUMMARY	status PARTIAL with reason CHARGE_LIMIT, DEADLINE, SOME_GAMES_FAILED or SOME_INPUTS_INVALID	Read the reason and the issue rows
Technical failure or retry	Dataset error rows, RUN_SUMMARY	errorCode, message; saved rows kept	Run the failed games again later

The RUN_SUMMARY record has the run status (COMPLETE, PARTIAL, NOT_FOUND, ERROR, INVALID_INPUT), the reason, counts, and one line per game with the number of reviews, why it stopped and newestReviewAt.

Output example

A review row (real, from the Counter-Strike 2 review list):

{
  "type": "review",
  "appId": "730",
  "gameName": "Counter-Strike 2",
  "reviewId": "219989924",
  "steamId": "76561198351194597",
  "recommended": false,
  "text": "The gameplay is nice but the game is infested with cheaters :(",
  "language": "english",
  "playtimeHours": 3097.1,
  "playtimeForeverHours": 3108.1,
  "playtimeLastTwoWeeksHours": 0,
  "votesUp": 940,
  "votesFunny": 57,
  "helpfulScore": 0.9106771349906921,
  "commentCount": 22,
  "steamPurchase": true,
  "receivedForFree": false,
  "refunded": false,
  "writtenDuringEarlyAccess": false,
  "primarilySteamDeck": false,
  "developerResponse": null,
  "developerRespondedAt": null,
  "createdAt": "2026-03-06T15:36:44.000Z",
  "updatedAt": "2026-03-06T15:36:44.000Z",
  "url": "https://steamcommunity.com/profiles/76561198351194597/recommended/730/",
  "storeUrl": "https://store.steampowered.com/app/730/",
  "source": "steam",
  "fetchedAt": "2026-10-17T11:57:22.670Z"
}

A game row (trimmed: categories and shortDescription shortened):

{
  "type": "game",
  "appId": "730",
  "gameName": "Counter-Strike 2",
  "appType": "game",
  "isFree": true,
  "price": 0,
  "initialPrice": 0,
  "currency": null,
  "discountPercent": null,
  "priceFormatted": "Free",
  "genres": ["Action", "Free To Play"],
  "categories": ["Multi-player", "Steam Workshop", "In-App Purchases"],
  "releaseDate": "Aug 21, 2012",
  "releaseDateIso": "2012-08-21",
  "comingSoon": false,
  "developers": ["Valve"],
  "publishers": ["Valve"],
  "platforms": ["windows", "linux"],
  "metacriticScore": null,
  "recommendationsTotal": 5315663,
  "reviewScore": 8,
  "reviewScoreDesc": "Very Positive",
  "totalPositive": 8517496,
  "totalNegative": 1410460,
  "totalReviews": 9927956,
  "positivePercent": 85.8,
  "requiredAge": 0,
  "shortDescription": "For over two decades, Counter-Strike has offered an elite competitive experience…",
  "headerImage": "https://shared.akamai.steamstatic.com/store_item_assets/steam/apps/730/…/header.jpg",
  "website": "http://counter-strike.net/",
  "countryCode": "us",
  "url": "https://store.steampowered.com/app/730/",
  "source": "steam",
  "fetchedAt": "2026-10-17T11:46:53.575Z"
}

null means Steam did not give a value (for example no Metacritic score, or no currency for a free game). 0 is a real zero. Review text is kept exactly as written, including Steam formatting tags like [b] or [spoiler]. Steam decides which reviews it shows: reviews that Steam hides (for example off-topic review bombs) or deletes are not returned.

Use case → fields

Buyer job	Input path	Output entity	Key fields	Next integration
Sentiment dataset	language, maxReviews	review rows	recommended, text	LLM or classifier
Patch monitoring	sortBy: recent, reviewsSince	review rows	createdAt, recommended	Daily chart
Price and score table	maxReviews: 0	game rows	price, positivePercent	Sheets or BI

👉 Try it now: open the input form

Pricing

Pricing unit

Pay per event: you pay for each saved review and game row, plus a small start and per-game fee. Apify platform usage is included.

Event	When	Price (USD)
Actor start	once per run; the fee is counted per GB of run memory (one fee at the default 256 MB)	0.001
Game processed	per app ID the run starts, also an unknown ID	0.0002
Review	per saved review	0.0004 Free · 0.00036 Bronze · 0.00032 Silver · 0.00028 Gold and above
Game details row	per game row (includeGameDetails: true)	0.001

A valid run with no matching reviews (a filter or a date with nothing new) costs the start fee plus $0.0002 per game, plus the game row if it is on. An unknown app ID costs only the $0.0002 game-processed fee. Issue rows (not_found, error) and entries that are not Steam app IDs are free. A failed game is never charged more than a successful one: you pay only for the rows that were saved.

Cost examples

Scenario	Input scope	Expected item/event count	Base charge	Optional charge	Stop/cap behavior
Prefill	1 game, 20 reviews	20 reviews + 1 game row	$0.0102 Free, $0.0078 Gold	game row included	stops at 20
One game in depth	1 game, 1,000 reviews	1,000 reviews + 1 game row	$0.40 Free, $0.28 Gold	game row included	stops at 1,000
Ten games	10 games, 100 reviews each	1,000 reviews + 10 game rows	$0.41 Free, $0.29 Gold	game rows included	stops at 100 per game
Game rows only	100 games, maxReviews: 0	100 game rows	$0.12	none	—
Unknown IDs	100 unknown app IDs	100 not_found rows	$0.021	none	NOT_FOUND

Set Maximum cost per run in the run options to cap a run: the Actor stops before a charge would go above it and returns what it already paid for (status PARTIAL, reason CHARGE_LIMIT).

👉 Try it now: open the input form

Limits

• Only the Steam store: no bundles, packages, DLC lists, workshop items or player profiles. Such links are skipped with an INVALID_INPUT row.
• Up to 100 games per run and up to 50,000 reviews per game in the input. In our tests 1,000 reviews took about 6 seconds and 5,551 reviews about 38 seconds (roughly 145 reviews per second), so at that rate the default 300-second timeout fits about 40,000 reviews (an estimate: our largest tested run saved 24,000 reviews per game); Steam rate limits can make it fewer. For more reviews, raise the run timeout or split the games over several runs.
• Steam limits the request rate. The Actor waits between requests and slows down on rate limits, so big runs take minutes, not seconds.
• If the run reaches its time limit, it stops about 15 seconds before the timeout and keeps all saved rows. The Apify run then shows FAILED, while RUN_SUMMARY says PARTIAL with reason DEADLINE when rows were saved (or ERROR when none were).
• If the run reaches your maximum charge, it stops, keeps all saved rows and ends as PARTIAL with reason CHARGE_LIMIT.
• If Steam fails for one game, other games continue; the failed game gets an error row and the run is PARTIAL with SOME_GAMES_FAILED.
• A reviewsSince date without a time means midnight UTC, and only reviews created strictly after it are returned. A reviewsUntil date without a time means the end of that day (23:59:59 UTC), and reviews created up to it are included.
• Steam shows brand-new reviews with a short delay (its own cache; in our tests about 10 minutes), so a review written just before a run may only appear in the next one.

👉 Try it now: open the input form

Scheduled runs: only new reviews

1. Run once with sortBy: recent.
2. In RUN_SUMMARY, copy newestReviewAt of the game.
3. In the schedule input, set reviewsSince to that time.

Only reviews created strictly after reviewsSince are returned, so the checkpoint review itself does not come back and a run with no new reviews returns no review rows. Steam itself filters by creation date in recent and helpful order, so such a run reads the game details and one review page per game; in updated order the Actor checks update times and stops at the first older one. Use newestReviewAt as a checkpoint only with sortBy: recent; in helpful order it is just the newest of the returned reviews.

A review that Steam shows late (see Limits) normally comes in the next scheduled run, because it is newer than the checkpoint.

👉 Try it now: open the input form

API, export and AI agents

• Export: the dataset downloads as JSON, CSV, Excel, XML or HTML. The Reviews view shows review columns (game and status rows appear there too; filter by type = review for a pure review table); GAMES gives the game table as JSON.
• API: start a run with POST https://api.apify.com/v2/acts/cliqtomedia~steam-reviews/runs and the input JSON; read rows from the run's dataset.
• Scheduling, webhooks, Make, Zapier, n8n: use the standard Apify integrations; the next system reads review rows by reviewId.

Use with AI agents (MCP)

• MCP server with only this Actor: https://mcp.apify.com?tools=cliqtomedia/steam-reviews
• Smallest useful input: {"appIds": ["730"], "maxReviews": 20}
• Key output fields: recommended (true = Recommended), text (review), playtimeHours (hours played at review time), votesUp (helpful votes), createdAt (UTC time). Game rows have price, reviewScoreDesc and positivePercent.
• The smallest run takes a few seconds and costs about $0.01; 1,000 reviews take about 6 seconds and cost $0.28–$0.40. Keep maxReviews small for quick answers.

👉 Try it now: open the input form

Troubleshooting

I got a not_found row

Symptom: no reviews and a not_found row. Likely reason: the ID is wrong, or the app has no store page (removed games, some tools and servers). Check: open store.steampowered.com/app/<id>/ in a browser. Action: fix the ID or use the store link.

I got an error row with INVALID_INPUT

Symptom: an error row with appId: null and the run is PARTIAL with SOME_INPUTS_INVALID. Likely reason: the entry is not a Steam app ID or a store link with /app/<id>/ (for example a bundle, package or profile link, or a typo). Check: the input and message of the row. Action: replace it with the app ID or the game's store link.

Fewer reviews than I asked for

Symptom: a game has fewer rows than maxReviews. Likely reason: the filters (language, reviewType, purchaseType, reviewsSince) leave fewer reviews, or Steam has no more. Check: stopReason for the game in RUN_SUMMARY (end_of_reviews, since_reached, max_reviews). Action: widen the filters.

No game row for a game

Symptom: reviews are there, the game row is missing. Likely reason: the game is not sold in countryCode. Check: detailsMissing in RUN_SUMMARY counts. Action: try us or another country.

The run ended with SOURCE_ERROR or SOME_GAMES_FAILED

Symptom: error rows for some games. Likely reason: Steam did not answer after several retries. Check: errorCode of the error rows; saved rows are kept. Action: run the failed games again later.

👉 Try it now: open the input form

FAQ

Do I need a Steam account or API key?

No. The Actor reads public Steam store data. No login, cookies or key.

Is personal data included?

Only the public steamId and the public review link. No names, avatars or profile details.

Are the numbers live?

Yes, each run reads Steam at that moment; fetchedAt shows when. Counters like votesUp and totalReviews change over time.

Why are the playtime numbers decimals?

Steam stores minutes. The Actor turns them into hours with one decimal place.

👉 Try it now: open the input form

Related Actors

• Google Play Reviews Scraper: Ratings, Replies & Versions — the same review analysis for a mobile game or app on Android.
• Apple App Store Reviews Scraper: Ratings & Review Text — reviews of the iOS version of a game.
• Reddit Scraper: Posts, Comments & Subreddits (No login) — what players say about the game in its subreddit.
• YouTube Comments Scraper: Replies & CSV Export — comments under trailers and gameplay videos.

👉 Try it now: open the input form

Support

For a bug, question or feature request, open an issue on the Actor's Issues tab. Add the run ID and your input. Never post tokens or private data.

Data comes from the official Steam Web API review service and the public Steam store (Valve Corporation). Your use of the data is subject to the Steam Web API Terms and the Steam Subscriber Agreement. Review texts belong to their authors. This Actor is not affiliated with or endorsed by Valve.

👉 Try it now: open the input form