Paste a city or category listing URL and get every restaurant on it back as rows: the name, its cuisines, the rating and review figure, the price band, a five-part address, an E.164 phone number, and links to the store and its image. Point it at one store instead and you get that single row.
Browse Uber Eats the way a customer would - a city, or a cuisine within a city - then hand us the address from the bar.
One row per restaurant, tagged with the query it came from and numbered in the order Uber Eats listed it. No residential proxy needed - the run goes through our shared pool.
Three input shapes, and two of them fan out: a city listing or a category-within-a-city listing returns every restaurant on it, while a store URL returns one. That is the difference between reading a market and reading a card.
Street, city, region, postcode and country come as five columns rather than one line to parse - and the phone number arrives in E.164, with a leading plus and no punctuation. Both are ready to load without cleaning.
Rows arrive numbered in the order Uber Eats served them, so what the platform pushed to the top of a city is recoverable after the fact. That ranking is the part a manual copy-paste always loses.
What an Uber Eats listing carries about a restaurant: who it is, what it cooks, how it scores, roughly what it costs, where it is and how to reach it.
Two of these columns behave in ways worth knowing before you build on them, so they are called out in the table rather than left for you to trip over. reviews is a display figure, not an exact count. Small restaurants come back with what look like real numbers, but past about a hundred the values arrive rounded - there is nothing between 1000 and 1500, or between 1500 and 2000. Summing that column, or averaging it, treats a rounded band as a measurement.
And price_range is not a reliable currency indicator. It arrives as one or two repeated currency symbols, but the symbol does not track the country column: restaurants with US addresses in Oklahoma came back carrying £ and € as well as $. Read it as a coarse price band if you use it at all, and never derive a currency from it.
Everything else is more straightforward. Every field is a string except position, so cast rating before you sort - and note that it is not zero-padded, so 4 and 4.2 both appear. cuisines is a comma-separated list inside one cell, and phone arrives in E.164 with a leading plus.
The export's header row, in order. Where a field is marked optional below, that describes what a restaurant published on its Uber Eats listing - not the scrape.
Famous Cajun/Chinese Grill - Memorial Rd.Chinese, Asian, Asian Fusion. Split on the comma if you want to group by a single tag.4 and 4.2 both occur, so cast before you sort or compare.£ and € as well as $. Optional - some listings carry none.OK, TX in our runs.Two traps, both easy to avoid once you know. reviews is a display figure rather than an exact count - past roughly a hundred it arrives rounded, so a sum or an average over that column is measuring Uber Eats' rounding, not its restaurants. And price_range is a band, not a currency: the symbol does not track country, and US-addressed rows came back with £ and € alongside $. Beyond that, every field is a string except position, which exists only in the JSON, so a CSV or XLSX export has fourteen columns rather than fifteen. rating is not zero-padded, cuisines is a comma-separated list in one cell, and phone is E.164.
Three shapes of Uber Eats URL, and two of them return the whole listing. All three come back in the same fourteen columns, one row per restaurant - and none of them needs a residential proxy.
A few examples of what a delivery listing is worth once it is a table rather than an endless scroll.
Start with the city URL to see everyone who delivers, then narrow to a category URL for the cuisine you care about. The five-part address means you can cut the result by postcode without writing a parser.
Rows come numbered in listing order, so you can keep the ranking alongside the rating and the price band. Re-run the same city later and the movement is the interesting part - not the snapshot.
Name, five-part address and an E.164 phone number arrive as their own columns, so a listing becomes a contactable list without cleaning. If you sell to restaurants - packaging, POS, ingredients - that is the shape you want.
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, file upload and every export format included.
Roughly $2 per 1,000 restaurants. The pre-flight estimator shows the row count and credit cost before a run starts - no surprise bills, no compute units to translate.
Volume pricing, dedicated workers and an SLA for mapping many cities on a schedule. Tell us your numbers and we will quote.
LIVESCRAPER10 at checkout.One platform is one slice of a delivery market. These three widen it.
The same job on Deliveroo, in a near-identical column set. Restaurants that appear on one and not the other are the gap worth knowing about - though note Deliveroo needs a residential proxy and this one does not.
The Dutch delivery platform, for the markets Uber Eats does not lead. What customers actually said, rather than the summary figure on a card.
Every restaurant in an area, not only the ones that deliver. It also returns an exact review count, which makes it a useful cross-check on the rounded figure you get here.
Short answer: these listings are public business information - which makes this one of the more straightforward pages on the site.
Everything collected here is shown on a public Uber Eats page to any visitor, signed in or not: the restaurant name, its cuisines, the rating and review figure, the price band, the address, the phone number and the header image. No login, no paywall, nothing behind an account, and no residential proxy required.
This export is about businesses, not people. There are no customer names, no review text, no order histories and no columns for any of them - the restaurant's phone number and address are the contact details it publishes precisely so customers can use them. That keeps this materially lighter on data-protection obligations than a reviews export. If you load these rows into a CRM for outreach, ordinary marketing rules still apply.
As always, Uber's own terms govern automated collection from their site - read them before you scale rather than relying on the fact that the pages are public. 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.
An Uber Eats city page is a delivery market in a scroll: every restaurant that delivers to an area, with its cuisines, its rating, its price band and its address, shown a card at a time. The Uber Eats Scraper turns that page into rows. You submit a city URL, a category-within-a-city URL or a single store URL - one per line, or as a CSV, XLSX, TXT or Parquet upload - and each restaurant comes back as a record with the name, a comma-separated cuisine list, the rating and review figure, the price band, a five-part address, an E.164 phone number and links to the store and its image. Fourteen columns in the file, fifteen in the JSON, one row per restaurant.
Two of those columns need a word of warning, and it is better said here than discovered halfway through an analysis. reviews is a display figure rather than an exact count: past roughly a hundred the values arrive rounded, with nothing between 1000 and 1500 or between 1500 and 2000, so summing or averaging that column measures Uber Eats' rounding rather than its restaurants. And price_range is not a currency - it arrives as one or two repeated symbols, and US-addressed rows came back carrying £ and € alongside $, so the symbol does not track the country column.
The rest is straightforward and ready to load. Every field is a string except position, so cast rating before sorting and note that it is not zero-padded - 4 and 4.2 both occur. The address arrives split into street, city, region, postcode and country, which is what makes a postcode-level view of a delivery market a sort rather than a parsing job, and the phone number arrives in E.164 with a leading plus and no punctuation. Rows are numbered in listing order, so what the platform pushed to the top of a city survives into the spreadsheet.
Unlike our Deliveroo scrapers, this one needs no residential proxy - the shared pool is enough. Start free: your first 500 rows cost nothing and need no credit card, and after that it is $0.002 per row, flat.