A Google Play scraper API returns Play Store listings as structured rows — title, developer and developer URL, category, description, rating and review count, price and currency, content rating, install band where Play shows one, icon and the localised store URL.
You pass package names (or store URLs) and it fetches one store page each, reading the page's schema.org block rather than its CSS. That choice is what makes the rows durable: Play reshuffles its interface regularly, but the structured-data block it emits for search engines stays stable. Packages that return nothing come back under failed, never billed and never invented.
$0.003 per delivered app, up to 50 apps per run. Nothing delivered means nothing charged.
Post an array of package names — com.spotify.music — or full Play URLs, up to 50 per run. Each becomes one store-page fetch through a residential exit in your chosen country, parsed from the embedded structured-data block. No browser.
This collector is search-less by design, and that is worth understanding rather than working around: Play's search results are heavily personalised and rendered client-side, so a "search Play" endpoint would return something that no real user saw. Package-keyed lookup returns exactly what the store publishes for that app. To discover packages in the first place, the search collector against site:play.google.com is the usual route.
installs is a band ("1,000,000+"), not a number, because that is all Google publishes. It is delivered as the string Play shows rather than converted into a false precision.
packages is required: an array of package names or store URLs. country and lang select the storefront and the metadata language — Play localises descriptions and prices, so both matter for comparative work. max_results caps how many packages the run actually fetches, which is a useful guard when the list comes from another job.
One row per app. installs is Play's band rather than an exact figure. rating and reviews are the current published values. developer_url links to the developer's other apps, which is the fastest way to map a publisher's portfolio. content_rating is the regional age rating and operating_system the requirement string where shown. description is the full store description, which is what ASO keyword analysis actually needs.
$0.003 per delivered app ($3 per 1,000). Nothing delivered means nothing charged, and the $2 monthly free credit covers roughly 666 apps here. Volume tiers take up to 30% off.
$0.003 per delivered app, $3 per 1,000 — about four times the App Store rate, and the reason is structural: Apple answers a whole result set in one response, while Play requires an individual page fetch per app.
Tracking 300 competitor apps weekly is $0.90 a week. Packages that do not resolve — delisted, region-locked, renamed — are returned in failed and are not billed, so an ageing package list costs coverage rather than money. The free $2 monthly allowance covers about 666 apps.
Google's Play Developer API is, like Apple's, a management tool for your own apps: releases, in-app products, subscriptions, your own reviews and reports. There is no endpoint that returns another developer's listing, so it does not compete with this — it does not address the same problem at all.
The widely-used open-source google-play-scraper libraries in Python and Node are the real alternative, and they do work. The trade is the usual one: they run from your IP, which limits throughput and invites blocks at volume; they parse Play's internal data structures, which move; and when they break, fixing them is your afternoon. This collector reads the structured-data block instead, runs through residential exits, and separates failures from results.
Specific things that bite:
Weekly ratings, review counts and install bands across a competitor set, with description changes as an ASO signal.
developer_url takes you from one app to everything the same publisher ships.
Run the same product against this and the App Store collector to compare how a product performs on each platform.
Full descriptions across a category show which terms competitors are actually targeting in store copy.
Up to 50 packages per run — lower than the App Store collector because each app is its own fetch. There is no search input: package-keyed lookup only, for the reason described above. Unresolvable packages return in failed without charge. Runs parallelise up to your plan's rate limit, 60 requests/minute on pay-as-you-go up to 1,200 on the top tier.
Legally: store listings are public marketing material and collecting public data is generally lawful in most jurisdictions, though Google's terms restrict automated access. Icons, screenshots and descriptions belong to the developer. Not legal advice.
There is a free allowance of $2 a month with no card, about 666 delivered apps at $0.003 each. The open-source google-play-scraper libraries are free and genuinely work — they run from your own IP, which is the constraint at volume, and they parse internal structures that move.
No, and that is deliberate. Play's search results are heavily personalised and rendered client-side, so a scraped 'search' would return something no real user saw. It works from package names. To discover packages, a site:play.google.com query through the search collector is the usual route.
$0.003 per delivered app — $3 per 1,000 — with no subscription. It costs more than the App Store collector because Play needs one page fetch per app while Apple answers a whole result set at once. Failed packages are never billed.
No — it returns Play's band, like '1,000,000+', because that is all Google publishes. Converting that to a number would invent precision Google deliberately withheld, so the string is delivered as shown.
They come back under failed rather than as rows of nulls, and are not billed. Setting country explicitly is how you distinguish a genuinely delisted app from one that is simply unavailable in the region your exit is in.
It reads the schema.org structured-data block Play emits for search engines rather than its CSS or internal data structures. That block changes far less often, which is why selector-based scrapers need regular attention and this does not.
It returns the rating and the review count, not the review text. Individual reviews are a separate surface with different volume and personal-data considerations.
Skip the marketing. QuantumProxies.io publishes a machine-readable site map, Markdown for every page, a free MCP server, and APIs billed from the same balance as your proxies.
Connect in one command: npx -y quantumproxies-mcp