A Google reviews scraper API returns the reviews of a single place as structured rows — rating, date in both display and ISO form, full review text, author name and profile, review images, and the owner's response with its date.
You can address the place three ways: by data_id from a Google Maps places run, by place_id, or by a plain name and location that the collector resolves for you. That last form makes ad-hoc lookups easy; the id forms are what you use in anything scheduled, because ids are stable and a name is not.
$0.0005 per delivered review, up to 100 reviews per run. Nothing delivered means nothing charged.
The collector opens the place's review feed and pages through it until max_results, up to 100 reviews per run.
sort_by is the input that decides what you are actually measuring. Newest gives you a monitoring feed. Relevance gives you Google's own idea of the representative reviews, which is what a prospective customer sees. Lowest rating is the complaint set — the fastest route to what is going wrong. Sorting is applied at the source, so a lowest-first run collects and bills the reviews you wanted rather than a hundred five-star ones you discard.
The owner-response fields are the ones people underuse. owner_response_text and owner_response_date let you measure whether a business replies to criticism and how quickly — a strong signal in competitive analysis, and the thing agencies most often want to show a prospect.
One of data_id, place_id or the query + location pair identifies the place. Prefer data_id in a pipeline: it comes straight from the Maps collector and cannot drift, whereas a name lookup can resolve to a different branch after a rebrand. sort_by selects newest, relevance, highest or lowest rating and applies at the source. max_results caps delivery and cost.
One row per review. text is the full review, not a snippet. date is Google's relative display ("2 months ago") and iso_date the parsed timestamp — use the second for anything time-series. author and author_url identify the reviewer, which is personal data and should be treated accordingly. images carries photo URLs where the reviewer attached them. owner_response_text and owner_response_date are the business's reply. place_name and data_id are stamped on every row so a multi-place collection stays a flat table.
$0.0005 per delivered review ($0.50 per 1,000). Nothing delivered means nothing charged, and the $2 monthly free credit covers roughly 4,000 reviews here. Volume tiers take up to 30% off.
$0.0005 per delivered review, 50 cents per 1,000 — one of the cheapest units in the catalogue, because review feeds are dense and predictable to read. Monitoring 200 locations weekly at 20 new reviews each is 4,000 rows, $2 a week.
Sorting at the source is the cost lever: a lowest-rating run of 20 reviews costs a cent and returns the complaints, rather than costing five cents to collect a hundred and discard eighty.
The Places API does return reviews, and it returns five of them. That is not a rate limit you can raise — it is the documented behaviour of the endpoint, and it is the single reason most review-analysis projects end up scraping. Five reviews will not support sentiment analysis, will not show a trend, and will not tell you what changed after a bad quarter.
The other constraint is the licence. Google's terms restrict how long Places content may be cached and what may be built from it, which sits awkwardly with keeping a review history to analyse over time. Reputation platforms exist partly to absorb that problem on your behalf.
If five current reviews and Google's caching terms suit your use case, the official API is simpler and you should use it. If you need the feed, this returns it — with sorting, owner responses and ISO dates.
The Maps review feed is a paginated, lazily-loaded panel, which creates predictable difficulties:
Sort by newest across your locations on a schedule and alert on anything below a threshold. iso_date makes 'new since last run' a trivial comparison.
Read competitors' lowest-rated reviews to find the operational failures customers actually complain about — the cheapest customer research available.
Full review text across hundreds of places is a real corpus. Five reviews per place is not.
Owner response fields let you measure whether a business or a competitor engages with criticism, and how fast.
Up to 100 reviews per run, one place per run. Monitoring many locations means one run each, parallel up to your plan's rate limit — 60 requests/minute on pay-as-you-go, up to 1,200 on the top tier. Sorting is applied at the source, so pick the sort that matches the question.
This is a personal-data collector and should be treated as one. Reviews are written by identifiable individuals, and author names and profile links are personal data under GDPR, CCPA and equivalents even though the reviews are public. You need a lawful basis to store them, individuals retain rights over them, and the review text itself is the author's copyrighted expression — aggregate analysis sits very differently from republication. Google's terms also restrict automated collection. Not legal advice.
There is a free allowance of $2 a month with no card, about 4,000 delivered reviews at $0.0005 each. Google's Places API is free within its credit tier but returns only five reviews per place, which is usually the reason people look for an alternative in the first place.
Because it returns five reviews per place. That is documented behaviour rather than a quota you can raise, and five reviews cannot support sentiment analysis or trend tracking. Its caching terms also limit how long you may keep what it returns.
$0.0005 per delivered review — 50 cents per 1,000 — with no subscription. Sorting is applied at the source, so a lowest-rating run collects and bills only the complaints instead of everything. Volume tiers take up to 30% off.
Three ways: data_id from a Google Maps places run, a place_id, or a name plus location that gets resolved to the first match. Use an id for anything scheduled — a name lookup can resolve to a different branch after a rebrand or a relocation.
Yes, the full text rather than the truncated preview. Scrapers that read the rendered page usually capture only the first couple of lines, because longer reviews render collapsed behind a More control.
Yes — sort_by set to lowest rating applies at the source, so the run collects and bills the low-rated reviews rather than everything. It is the cheapest way to find out what customers actually complain about, yours or a competitor's.
Reviews are publicly displayed, but they are written by identifiable people, so author names and profile links are personal data under GDPR and similar laws and need a lawful basis to store. The text is the author's copyrighted expression, so aggregate analysis and republication are different propositions. Google's terms also restrict automated access. Take proper advice.
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