Point it at a zillow.com/profile/ page and every review on it comes back as a row - who wrote it, the rating they gave, the date, the headline, the review itself and how it is classified.
This is the third of our three Zillow profile services, and they all take the same input. The agents scraper gives you one row per agent, the transactions scraper one row per deal, and this one gives you one row per review.
https://www.zillow.com/profile/agent-name/ - one per line, or upload a CSV, XLSX, TXT or Parquet file.One row per review, tagged with the profile URL it came from. Pages are fetched through our proxy pool rather than your real IP - but the pool alone is not enough here, which the next section explains.
A star average tells you very little about why. This export carries the headline and the full review on every row, so you can read what clients actually said - and run the text through whatever analysis you already use.
Alone among our three Zillow profile services, this one lets you order the run: Newest first, Rating low to high or Rating high to low, plus a separate From (oldest first) toggle. Combined with a per-query limit, that is how you take the most recent reviews rather than an arbitrary slice.
The same profile URLs work on all three of our Zillow services. Run them through each and join on the profile URL and you have the agent's summary, the deals they transacted and the reviews clients left, side by side.
This is the first thing to know about this scraper, so it is not buried further down the page.
Zillow blocks datacenter IPs, and PerimeterX is the mechanism. A run without a residential proxy does not error, does not warn and does not stop - it finishes successfully and hands you a file with the seven-column header row and nothing under it. That is a worse failure than a crash, and on this page in particular, because an empty file reads as "this agent has no reviews" - which is a conclusion you might well act on. If a profile you know is reviewed comes back empty, check the proxy first.
We can say that with some confidence rather than as a disclaimer. Every run export we hold for this service is empty, and so is every one for our other two Zillow profile scrapers - the agents and transactions services - across runs spread over several months. Every one carries a correct header row and zero data rows. That is a consistent, well-corroborated picture of what happens without a residential exit, and it is why the requirement is in the hero rather than in a footnote.
The consequence for this page is that the column names are solid and the cell contents are not. The seven names below, and their order, are confirmed twice over - by the identical XLSX header row of four separate runs and by the column array the platform publishes for this service, which agree exactly. What we will not do is tell you what the values look like, because we have not seen one. Where a description below would have needed a format, a scale or a vocabulary, it says so instead of guessing.
The export's header row, in order. Descriptions say what each field is for; they deliberately claim nothing about its format, scale or vocabulary - see the note above.
max_rating - so the scale is not recoverable from the file itself. Confirm it against a real run before comparing these against another platform.The names are solid; the contents are not yet. The seven columns above and their order are corroborated by two independent sources - the identical XLSX header row of four separate run exports and the column array the platform publishes for this service - so build your schema around those names with confidence. What this page will not tell you is what the cells hold: every export we hold is a header with no rows beneath it. Note the one structural gap: there is no max_rating column here, so the rating's scale is not recoverable from the file and needs confirming against a real run. And note author, which is a reviewer's name and makes this the one Zillow export of the three that is squarely about members of the public. Zillow blocks datacenter IPs, so a live run needs a residential proxy - that is the reason those exports are empty, not any doubt about the column set.
The same URL as our other two Zillow services, plus the ordering controls neither of them has. Set them deliberately - with a per-query limit they decide which reviews you get, not just what order they arrive in.
A few examples of what reviews are worth once they are rows rather than a panel you scroll through one agent at a time.
Sort by rating low to high, take the bottom slice with a per-query limit, and you have the criticism without wading through the praise. The full review text is on every row, so the themes are yours to pull out however you normally would.
A profile shows a star average; this shows the reviews behind it. Newest first with a small limit gives you the recent picture rather than the flattering all-time one, which is usually the question you are actually asking.
Run one list of profile URLs through all three of our Zillow services and join on the profile URL. The agents export gives the headline figures, the transactions export the deals, and this one what clients wrote - three views assembled from a single list.
No subscription, no minimum, no recurring bill. Your first 500 rows are on us - after that, pay-as-you-go at the same flat rate as every other scraper here.
Every new account, one-time. No credit card required. Per-query limits, sorting, file upload and every export format included. The residential proxy this scraper needs is yours to supply, on the free tier and after it.
The same flat rate as every other scraper on the platform. With one row per review, the sort and the per-query limit are what decide how many rows an agent costs you - set them before a large run, not after. The estimator shows the cost before it starts.
Volume pricing, dedicated workers and an SLA for tracking reviews across many agents on a schedule. Tell us your numbers and we will quote.
LIVESCRAPER10 at checkout.This one returns the reviews. Its two siblings take exactly the same input and return something else - and the third link is for when you want the same question answered off Zillow.
One row per agent instead of one per review - name, brokerage, phone, website and the profile's own rating and review count. That is where the summary figures live; this is where the opinions behind them do.
One row per deal listed on the same profile. Useful next to this one when a review's tone and an agent's actual transaction history seem to be telling different stories.
Reviews of the same brokerage from a different audience. A second platform is the usual way to tell whether a pattern is real or an artefact of who happens to review on Zillow.
Short answer: the reviews are published to be read - but this export names the people who wrote them, and that is a bigger obligation than either of our other two Zillow services carries.
Reviews on a Zillow agent profile are shown to anyone who opens the page, signed in or not. No login is used, no paywall is crossed and no account is touched, and the run is fetched through our proxy pool rather than your own IP. Collecting publicly visible feedback for research is long-established practice.
The difference here is author. Our kununu export makes a point of carrying no reviewer identity at all; this one carries a name, and that name belongs to a member of the public rather than to a professional acting in a business capacity. A person's name attached to their written opinion is personal data at the more sensitive end of the range - an opinion can reveal a good deal about someone. If you are in the EU or the UK, running this makes you the controller: you need a lawful basis, and because the data did not come from the person, you have to be able to tell them you hold it and where it came from. Do not put this column into a marketing list.
Two more points worth stating plainly. Zillow's terms of use restrict automated access to their site, and the datacenter-IP blocking is there because they mean it - "the page is public" is an argument about the data, not about the contract you accept by using the site, so read those terms before you scale. And the review column is somebody else's writing: collecting it for analysis is one question, republishing it verbatim is a different one, and worth checking before you do it. We run no third-party trackers on the data layer, and your exports auto-delete after 30 days.
The questions we hear most. Anything else? Talk to us - humans, not bots, write the answers.
Your first 500 rows are free - no card, no subscription. After that it is $0.002 per row, flat.
A Zillow agent profile shows a star average and, beneath it, the reviews clients have written. The Zillow Reviews Scraper turns those into rows. The input is a zillow.com/profile/ URL, pasted one per line or uploaded as a CSV, XLSX, TXT or Parquet file, and every review comes back as a record with the author, the rating they gave, the date, the headline, the review text and a classification column. Seven columns, one row per review, each tagged with the profile URL it came from so a file covering many agents stays separable.
This is the third of our three Zillow profile services, and all three take the same input. The agents scraper returns one row per agent - the profile summary, with contact details and the profile's own rating and review count. The transactions scraper returns one row per deal. This one returns one row per review. Run a single list of URLs through all three and join on the profile URL, and you have the headline figures, the deals and the opinions assembled together. This one is also the only one with sort controls: a Sorting select offering Newest first, Rating low to high and Rating high to low, plus a separate From (oldest first) toggle. With a per-query limit those controls decide which reviews you get, not merely their order.
One thing needs saying before you plan a run. Zillow blocks datacenter IPs, and a run without a residential proxy returns nothing. Worse, it does not tell you: the job finishes successfully and hands you the seven-column header row with no data under it - which on this service reads as an agent nobody has reviewed. Every export we hold for this service is in that state, as is every one for our other two Zillow profile scrapers, across runs spread over several months, which is why the requirement is stated at the top of this page rather than in a footnote. It also means the column names are confirmed and the cell contents are not: we will not tell you what format date arrives in, what vocabulary review_type uses, or what scale rating is on - and on that last one, note that there is no max_rating column, so the scale genuinely cannot be recovered from the file.
On the legal side, this export deserves more care than its two siblings. The reviews are public and collecting published feedback for research is ordinary - but author names a member of the public rather than a professional, and a person's name attached to their written opinion is personal data. In the EU and the UK that makes you the controller, with a lawful basis to establish and an obligation to say where the data came from. Zillow's terms restrict automated access on top of that, and the review text is somebody else's writing, so republishing it verbatim is a separate question from analysing it. Start free: your first 500 rows cost nothing and need no credit card, and after that it is $0.002 per row, flat, with the residential proxy as a separate cost.