Steam Review and Market Scraper
Steam pages mix server-rendered markup with JavaScript-driven widgets, so a plain HTTP fetch often returns an incomplete page.
This API takes a Steam URL, runs it in a real browser, rotates the outbound IP and returns the rendered result. You can then read the HTML yourself or ask for structured JSON via extraction rules. The sections below cover what teams collect, how the request options behave, what each option costs in credits, and where the plan limits sit.
What teams collect from Steam
Three data families dominate Steam collection work: user reviews, store pricing, and player-activity statistics. Each has a different refresh cadence and a different tolerance for partial pages.
- Reviews. Review text, helpfulness votes, playtime at review time and review timestamps. Useful for sentiment scoring and for tracking how a game's reception shifts after patches or sales.
- Prices. Store price and discount state per game. Price work is sensitive to region, because Steam shows different currencies and discount schedules depending on where the request originates.
- Player statistics. Current player counts, concurrent-user peaks and achievement statistics. Good for popularity tracking and for comparing titles in the same genre.
A single game page can supply all three, which is why one rendered request per game per refresh cycle is usually enough.
Request options and what they change
The base call is a URL plus your API key. Everything else is optional and changes either the fidelity of the response or the credit cost.
- render_js — executes the page in a browser so JavaScript-built sections appear. Steam review widgets and some price blocks depend on this.
- stealth_proxy — routes through a stealth proxy tier intended for pages that challenge automated traffic.
- country_code — routes the request through a residential IP in the named country, which matters when Steam serves region-specific pricing or content.
- extract_rules — CSS (or XPath) selectors mapped to field names; the response comes back as parsed JSON.
- ai_extract_rules — a natural-language description of the fields you want; no selectors required.
- Screenshots — full-page capture, useful for visual verification of a price or banner state.
- Markdown output — the page returned as clean Markdown for downstream LLM use.
Pick the smallest set of options that produces a complete page. Every option you add raises the per-request credit cost.
Credit cost per request
Credits are consumed per successful request, and the amount depends on which options are active. The published cost rules are:
- Base request: 1 credit.
- JavaScript rendering enabled: 5 credits per successful request.
- Premium proxies: 10 credits with JS rendering disabled, 25 credits with JS rendering enabled.
- Stealth proxy: 75 credits.
So a plain fetch of a Steam page costs 1 credit, a rendered fetch costs 5, a rendered fetch through premium proxies costs 25, and a stealth-proxy fetch costs 75. The product page states that credits are valid for one month and do not roll over. Plan sizes and concurrency limits appear in the plan comparison below; those figures are a dated reference from the product page, not a current offer, so confirm terms before committing to a volume.
Worked example: one game, price plus reviews
Suppose you want a daily snapshot of one game's store price and its 20 most recent reviews, and the review section only appears after JavaScript runs.
- Send the store URL with render_js enabled and extract_rules mapping the price field and the review list. That is one rendered request at 5 credits.
- If the store serves a different price in the target market, add country_code and repeat. Each additional country is another 5-credit request.
- If the page starts challenging the request, escalate to premium proxies (25 credits with JS) rather than retrying the cheap call repeatedly — retries on a blocked page still consume credits without producing data.
A 30-day run at one rendered request per day is 150 credits. Adding three country variants per day brings it to 600 credits for the month. Compare that against the plan credit allowances in the pricing table before choosing a tier.
Three different failures, three different fixes
Teams often treat every empty response as the same problem. They are not.
- Target-side throttling. Steam is rate-limiting aggressive patterns. The fix is to space requests out or spread them across sessions; the API itself does not impose per-target rate limits, only per-plan concurrency limits.
- Plan concurrency limit. You have hit the number of simultaneous requests your plan allows. The fix is to queue work or move to a plan with higher concurrency — the pricing table lists the concurrent-request figure for each tier.
- Bot challenge on the page. The response is a challenge page rather than the game page. The fix is to enable stealth_proxy or premium proxies, accepting the higher credit cost.
Diagnosing which of the three you are hitting before changing options avoids paying for stealth proxies on what is really a concurrency problem.
Field map: data type to request option
| You want | Enable | Notes |
|---|---|---|
| Review text, votes, playtime, timestamps | render_js + extract_rules or ai_extract_rules | Review widgets are JS-driven |
| Store price in a specific market | country_code | Prices vary by region |
| Current player count, concurrent peaks | render_js | Stats blocks load dynamically |
| Achievement statistics | render_js + extract_rules | Map each stat to a field name |
| Visual proof of a price or banner | Screenshot option | Full-page capture |
| Page text for an LLM pipeline | Markdown output | No selectors needed |
| Structured JSON without writing selectors | ai_extract_rules | Natural-language spec |
Setup path
- Create an account to get an API key. The product page advertises 1,000 free credits with no card required; confirm the current amount on the signup flow.
- Install an SDK for your stack — Python, Node.js, Go, PHP, Ruby and Java are published, and plain cURL works for any language.
- Send the Steam URL with your key. Add render_js for JS-driven sections, country_code for regional pricing, and extract_rules or ai_extract_rules when you want JSON rather than HTML.
- Log the credit cost per request so your monthly burn is predictable against the plan allowance.
Documentation examples exist in Python, Node.js, cURL and a request builder, so the first working call is usually a copy-and-key-swap.
| Plan | Price /mo | API credits | Concurrent requests | JS rendering | Rotating & premium proxies | Geotargeting | Screenshots & extraction rules | Priority email support | Dedicated account manager | Team management |
|---|---|---|---|---|---|---|---|---|---|---|
| Hobby | $19 | 75,000 | 25 | Included | Included | Included | Included | Not included | Not included | Not included |
| Freelance | $49 | 250,000 | 50 | Included | Included | Included | Included | Not included | Not included | Not included |
| Startup (Recommended) | $99 | 1,000,000 | 100 | Included | Included | Included | Included | Not included | Not included | Not included |
| Business | $249 | 3,000,000 | 200 | Included | Included | Included | Included | Included | Included | Included |
| Business + | $599 | 8,000,000 | 400 | Included | Included | Included | Included | Included | Included | Included |
| Enterprise 14 | $999 | 14,000,000 | 500 | Included | Included | Included | Included | Included | Included | Included |
| Enterprise 24 | $1,599 | 24,000,000 | 600 | Included | Included | Included | Included | Included | Included | Included |
| Enterprise 41 (Recommended) | $2,399 | 41,000,000 | 700 | Included | Included | Included | Included | Included | Included | Included |
| Enterprise 69 | $3,599 | 69,000,000 | 800 | Included | Included | Included | Included | Included | Included | Included |
| Enterprise 120 | $5,799 | 120,000,000 | 900 | Included | Included | Included | Included | Included | Included | Included |
Prices exclude VAT. The product page states credits are valid for one month and do not roll over. These are the product page's published figures, not a current offer; confirm pricing and allowances before purchase.
Frequently asked questions
Do I need an account to scrape Steam?
You do not need a Steam account. You do need an API key from this service: sign up, receive the key and a trial credit allowance, then send the Steam URL. The service handles rendering and proxy rotation on its side.
Can I scrape Steam from a specific country?
Yes. Add country_code to the request and it routes through a residential IP in that country. This is the lever for region-specific pricing and content, and each country variant is a separate request with its own credit cost.
What's the rate limit for scraping Steam?
There is no per-target rate limit. Limits are on concurrency per plan — the pricing table lists the simultaneous-request figure for each tier. If Steam itself throttles aggressive patterns, space requests out or spread them across sessions.
How do I get structured JSON from Steam?
Two routes. Pass extract_rules with CSS selectors mapped to field names, or pass ai_extract_rules with a natural-language description of the fields. Either way the response returns as parsed JSON.
Which programming languages can I use to scrape Steam?
Any language that can make an HTTP request. Published SDKs cover Python, Node.js, Go, PHP, Ruby and Java, and plain cURL works everywhere else.
How much does it cost to scrape Steam?
Per-request credit cost depends on options: 1 credit base, 5 with JavaScript rendering, 10 or 25 with premium proxies (without or with JS), and 75 with stealth proxy. Plan prices and credit allowances appear in the plan comparison above; those are the product page's published figures, a dated reference rather than a current offer, so check terms before purchase.