A Udemy course carries two prices at once, and the cheaper one expires in a day or two. A Udemy scraper that runs once tells you nothing; a feed that runs daily tells you how the whole catalog is priced.
ScrapeIt writes the collector, runs it on our machines and repairs it when the platform moves. Anti-bot handling, proxy rotation and CAPTCHA solving are part of the managed service, and they earn their keep here: the site puts a browser check in front of the markup, and no single address holds a steady rate across a catalog this size.
We plan collection the way the site invites it, from the sitemap and from category, subcategory and topic listings, not from faceted search robots.txt closes. Course pages, curriculum trees and review pages run on separate cadences: a price changes weekly, a syllabus does not. A sample extract comes first, so field names and coverage can be checked before the full run. After sign-off we run to your schedule and watch for layout changes so the feed does not go stale. Delivery is CSV, JSON, XLSX or an endpoint you call, pushed to S3, SFTP or your warehouse.
A Udemy course page is denser than it looks. These are the fields we extract Udemy data into, in the platform's own names.
Udemy runs one catalog behind nineteen interface locales, as path prefixes such as /de/, /ja/ and /zh-tw/. Currency follows the storefront, and so does number formatting, which is where feeds quietly break: the same course prints 474,422 ratings in one block and 474.422 in another on one page. Enum fields localise too, so a record pulled from a German storefront returns the level as Einsteiger:innen, not Beginner, and a free course prices as Kostenlos. Anything you group by has to be normalised on the way in.
Language is two separate facts. Teaching language is one value per course, and the filter lists roughly thirty-seven of them, from English down to single-course entries in Georgian and Uzbek. Subtitles are a second list, and the filter splits machine tracks from human closed captions with an explicit CC suffix, so a course can advertise thirty subtitle languages and hold one real caption track.
Udemy Business is a separate storefront, not a second catalog: its sitemap file holds a single demo request page, so the seat-based library is not public. What is public is the collection flag on each course. Plan pages put the subscription collections at close to thirty thousand curated courses against a marketplace of more than 250,000. Around them sit further surfaces: certification voucher listings, a certification browse hub, thirteen Career Accelerator tracks, and topic pages carrying learner totals, course counts and an average rating per subject. Instructors and review authors we treat as a data type, taking the display name Udemy publishes and building no named profiles or contact lists.
Udemy sorts its catalog on three levels, and a Udemy scraper has to respect all three. Thirteen subject categories sit at the top, among them Development, Business, IT and Software, Design and Marketing. Under them run 128 subcategories, from Web Development to Esoteric Practices and Money Management Tools. A fourteenth branch is not a subject at all: /courses/free/ is a price bucket split into Essential Tech, Personal Growth and Wellness, and Productivity and Professional Skills. Below the subcategory sits the topic, and topics are what the filters run on, each carrying a numeric label ID: 8322 is Web Development, 6368 is JavaScript, 7380 is Python.
Addresses are short and stable. A course lives at /course/ plus its published title and carries a numeric ID that reappears in the enrollment path, the mobile deep link and the CDN image name. Instructor pages answer to two forms, /user/ plus a handle and /user/ plus a UUID. Topics live at /topic/ plus a slug, with a free-only variant ending in /free/, and single preview lessons get their own pages under /tutorial/.
The sitemap index is the way in. It lists 34 files: navigation, category, subcategory, sixteen topic files, three instructor files, seven course files, preset topics, Udemy Business, aggregators and certification vouchers. Each course file holds up to fifty thousand addresses with a lastmod date, roughly 334,000 course URLs in all in September 2026. robots.txt then sets the rules: any URL with a query string is closed, with one exception for the ?p= page parameter, so pagination is open and faceted search is not. Nineteen interface locales exist as path prefixes, but a locale course page points its canonical at the plain English address, so one course stays one record instead of nineteen.
Get a Quote
Developers
Customers worldwide
Pages extracted
Hours saved for our clients
€199 / one-time
setup fee - included
€169 / mo
setup fee €499
€229 / mo
setup fee €499
€349 / mo
setup fee €499
€549 / mo
setup fee €799
Every paid course carries two prices at the same moment, and Udemy price monitoring only makes sense once you record both. list_price is the tier the instructor set. price is what the buyer pays today. Between them sits a discount object with saving_price, has_discount_saving and discount_percent, and under that a campaign block naming the code that produced the cut, its start_time and end_time, whether it is public, whether the instructor created it, and how many uses remain. On a bestselling course the gap runs to 86 percent off, the campaign lasts a handful of days, and the checkout link already carries the code.
That is why a single pull is worthless. The deal window closes, a new campaign code opens, and the same course sits at a different number. A stored Udemy data feed turns that churn into a series: how deep the cut goes by category, how long a campaign runs, which instructors opt into platform deals and which hold their list price.
Three more prices hide behind the first two. The record keeps a list price string in US dollars next to the local currency amount, so a course reading 139.99 euro on the page still carries a 199.99 dollar tier underneath. Separate Apple and Google in-app price fields quote a third, lower figure for buying inside the mobile apps. Subscription changes the maths again, because a course in the Personal Plan collection is reachable for a monthly fee rather than a purchase. Free courses close the set: is_paid turns false, the price string localises, certificates switch off and the feature list empties.
Learn how to use web scraping to solve data problems for your organization
Leveraging advances in technology, the AI-powered web scraper has skyrocketed in demand and is helping to expand capabilities by automating tedious daily tasks and speeding up data collection from thousands of websites several times over.
Web scraping is a method of obtaining web data by extracting it from pages of web resources with the help of a program, that is, in automatic mode. It is used to syntactically convert web pages into more usable forms.
If you specialize in machine learning, you need to feed large amounts of data to the algorithms. Web scraping is the easiest and the most efficient method of collecting the data from all over the Internet.
ScrapeIt is a managed data collection agency. We write the crawlers, run them on our machines, fix them when a site changes and hand over clean files. You do not maintain code and you do not manage proxies.
We also say no. If a smaller scope answers your question, you hear that on the first call. We do not sign in as a student to reach paid lecture video, we do not touch Udemy Business seats, and we do not assemble personal profiles of instructors or reviewers. What is public we collect well; what is not, we tell you about before anyone signs.
Not a usable one for catalog work. The Udemy Affiliate API v2.0 still has a docs page, but it is marked deprecated and access was discontinued on 1 January 2025; the affiliate program now runs on the Impact network and gives tracking and deep links, not a data feed. The Instructor API v1.0 is live at instructor-api/v1 with a bearer token, but it reaches only courses you teach, plus their questions, reviews and message threads. It is throttled to 100 requests per 10 seconds, caps page_size at 100, and its own overview calls it maintenance mode. A per-course read endpoint answers without a key, but nothing public lists or searches courses, which is the gap a Udemy scraper fills.
Yes, with one hard ceiling worth knowing before you budget. The review listing stops at ten thousand items per course no matter how many ratings the course holds, and it is ordered newest first, so a course with 474,422 ratings yields its most recent ten thousand and nothing older. Most of those entries are a star rating with no text at all; written reviews are the minority. Each one carries a rating on the half-star scale, a created and modified timestamp, the review body and the reviewer's public display name. We deliver ratings, dates and text, and leave named private individuals and contact details out by default.
Daily at minimum for anything price-driven. Udemy discount campaigns are short: a campaign block on a live course typically spans a few days, the site counts the remaining time down in the buy box, and a new code takes over when the old one expires. Refresh weekly and you will record only a handful of the year's price points and miss the depth of most cuts. Ratings, enrolment counts and the last-updated date move slowly enough for a weekly pass, and the curriculum tree can be refreshed monthly. We usually split a Udemy project into those three cadences so you are not paying to re-collect a syllabus that has not changed.
Yes, and it needs to be asked for explicitly, because a single collection run sees one storefront. Udemy serves nineteen interface locales as path prefixes and quotes the price in the currency of the market it thinks you are in, so a course reads one figure in euro and another in dollars. Teaching language is a separate field from subtitle tracks, and subtitle tracks are split between machine translations and human closed captions. We set the target markets up front, normalise currency, level and decimal formatting into one schema, and keep the original strings in a raw column so nothing is lost in the conversion.
Udemy's terms of use restrict automated access to the platform, so this is a scoping question, not a checkbox. We work only with pages Udemy publishes to anonymous visitors, follow the robots.txt rules on query strings and pagination, and we do not sign in, share accounts or reach behind a purchase to lecture video, quizzes or Udemy Business seats. Course authors and review authors are collected as a data type: the display name Udemy itself publishes, nothing appended from elsewhere, and no contact lists. Where a requirement sits outside that line we say so before the project starts, and the compliance call stays with you and your counsel.
Step 1 - Make a Request
You share your needs, expectations, and desired timeframe. We’ll suggest the best solution based on your request and budget.
Step 2 - Configuring Custom Web Crawlers
Our specialists configure the crawlers and extract a sample dataset for your review before proceeding with the full-scale extraction.
Step 3 - Collect and Deliver
Once you approve the sample, we launch the project and start full data collection. We gather, filter, and structure the data for easy use, delivering it on time in your preferred format.
Step 4 - Maintain and Support
Our team manages ongoing processes, monitors website changes, and supports all data extraction cycles. We can also help integrate data into your systems or create dashboards to simplify analysis.
Scrapeit Sp. z o.o.
10/208 Legionowa str., 15-099, Bialystok, Poland
NIP: 5423457175
REGON: 523384582
