Guide

Web Scraping in the Netherlands: What the AP Guidance Means

The Dutch data protection authority once called scraping almost always illegal. Learn what its revised April 2025 guidance says, what changed after the KNLTB ruling, and what it means for teams that collect data from Dutch websites.

Narmin Kamilsoy
Narmin Kamilsoy Author
13 min read
Web Scraping in the Netherlands: What the AP Guidance Means

On 1 May 2024, the Autoriteit Persoonsgegevens (AP), the Dutch data protection authority, published guidance on scraping by private organisations and individuals. The AP warned that scraping would almost always be a breach of the GDPR, and Dutch media headlines summed it up as "bijna altijd illegaal", almost always illegal. In April 2025, the AP published a revised edition of the same guidance, after the Court of Justice of the European Union rejected the narrow reading of legitimate interest that the AP had argued for in a Dutch case.

This guide explains what the current edition says, what changed, and what it means in practice for teams that collect data from Dutch websites, including through proxies. It describes the AP's guidance and the sources behind it. It is not legal advice.

1
1 May 2024
The AP publishes its scraping guidance.
→
2
4 Oct 2024
The CJEU rules in KNLTB that a commercial interest can be a legitimate interest, provided it is lawful.
→
3
April 2025
The AP publishes a revised edition citing that ruling.
→
4
7 Jul 2026
The EDPB adopts draft guidelines on scraping for generative AI.

What the AP Guidance Covers

Scope

The guidance is about the GDPR, known in Dutch as the AVG, and only for private organisations and individuals. It does not cover scraping by government bodies. It also states that it does not deal with other rules, such as copyright or the terms that websites set for their use, and notes that the AP does not supervise those rules.

It defines scraping as the automated collection and storage of information from web pages. Web crawling is included: according to the guidance, a fixed list of URLs is scraping, while a list that is adjusted dynamically during the process is crawling.

When the GDPR applies

The GDPR applies when personal data is processed and the processing falls within its material and territorial scope, subject to limited exceptions. The guidance adds that when you scrape information from the internet, you almost always also scrape personal data. According to the guidance, a legal basis is needed even when the information is publicly accessible, and even when you delete the personal data immediately after collecting it.

Two exceptions are discussed. The household exemption covers individuals who scrape for purely personal use and share the data only with a small group such as family or friends. The guidance warns against publishing the data, including in a public repository on GitHub. The second point is territorial: organisations outside the EU can also be covered by the GDPR, and according to the guidance, an organisation that scrapes personal data from .nl or .eu websites will likely collect data about people in the EU. It advises applying geographic filters to the URLs you visit, not to the data after collection, while noting that this is not watertight, because many Europeans are also active on international .com websites.

From "Almost Always Illegal" to the April 2025 Edition

The May 2024 edition

The first edition took a strict line on the legal basis that private organisations usually depend on: legitimate interest. In the AP's view at the time, an organisation with only a purely commercial interest could not successfully rely on that basis. The AP chair, Aleid Wolfsen, also stressed that information being public does not mean consent: "dat informatie over jou openbaar is, betekent niet automatisch dat jij ook toestemming geeft voor scraping."

The KNLTB ruling

On 4 October 2024, the Court of Justice of the European Union ruled in case C-621/22, a Dutch case involving the Royal Dutch Lawn Tennis Association (KNLTB). The Court held that a controller's commercial interest "could constitute a legitimate interest" under the GDPR, "provided that it is not contrary to the law." Days later, the AP told Security.NL that it would adjust its interpretation on this point. In November 2024, the Dutch IT industry association NLdigital asked the AP to withdraw the scraping guidance until it was revised.

What the April 2025 edition says

The April 2025 edition no longer excludes purely commercial interests. Citing the KNLTB ruling, it states that an interest must be lawful, that it does not have to be laid down in law, and that it must not be contrary to the law. Its examples of legitimate interests include fraud prevention, improving the security of computer systems, and the freedom of information in Article 11 of the EU Charter of Fundamental Rights. The summary of its findings also softened. The 2024 edition said that only (very) targeted scraping, in Dutch "(zeer) gerichte scraping", seemed capable of being lawful for private organisations. The 2025 edition says that mainly targeted forms of scraping by private organisations can be lawful.

The main changes are summarised below:

Topic May 2024 edition April 2025 edition
Commercial interests A purely commercial interest cannot support a successful reliance on legitimate interest, in the AP's view The interest must be lawful and not contrary to the law; it does not have to be laid down in law (citing the CJEU KNLTB ruling). Commercial interests are no longer singled out
Overall conclusion Only (very) targeted scraping seems capable of being lawful for private organisations Mainly targeted forms of scraping by private organisations can be lawful
Examples of more feasible scraping News websites, own web pages with reviews, security forums Same examples

What did not change is the overall message: the guidance still finds that the features of scraping often make it difficult or even impossible to meet the conditions for legitimate interest.

Newer European guidance

On 7 July 2026, the European Data Protection Board adopted Guidelines 03/2026 on web scraping in the context of generative AI, version 1.0. The public consultation runs from 8 July to 30 October 2026. They deal with scraping personal data to develop and train generative AI models, not with scraping in general, and may still change after the consultation. Like the AP guidance, they state that publicly posted data does not amount to consent, and that the absence of a robots.txt file does not either.

Legitimate Interest: How the AP Assesses Scraping

Why other legal bases rarely work

According to the guidance, legitimate interest will usually be the only legal basis available to private organisations. Consent is possible in principle, but identifying and asking everyone in advance is often difficult or even impossible. Publishing information online is not consent to scrape it: the guidance says people who make their data visible to everyone can be seen as allowing others to view it, but not to scrape or otherwise process it. Scraping is also treated as a new processing operation, not as compatible further use of data that is already online.

Three cumulative conditions

Legitimate interest requires all three of these conditions:

There is a legitimate interest of the organisation or a third party.
The processing is necessary for that interest.
The interests, fundamental rights, and freedoms of the people whose data is processed do not override that interest.

Factors the AP weighs

For the second and third conditions, the guidance lists features of scraping that you have to take into account:

Scale: the broader the scraper searches, the greater the intrusion. A database that can be searched by individual person increases it considerably.
Sensitive data, such as location or financial data.
Expectations: even people who publish their own data do not always have to expect that it will be processed for another purpose.
Consequences: a statistical analysis whose results cannot be traced back to individuals, or a sentiment analysis of how a new product is received, has smaller consequences than profiling people or deciding whether to hire someone based on their posts on social media and online forums.
The weak position of the people concerned, who often do not know their data is being scraped.

Safeguards

Additional safeguards, beyond what the GDPR already requires, can tip the balance, although the guidance says they will certainly not always lead to a successful reliance on legitimate interest. Its examples include:

Measures for extra transparency, for example information about the scraping on the websites being scraped, or a statement on the client's own website that it uses scraping.
Deleting, anonymising, or pseudonymising personal data as quickly as possible.
Applying the right to erasure more broadly than Article 17 GDPR requires, for example by always honouring erasure requests.
Following internet standards such as robots.txt.

The guidance adds that a website not using such standards does not mean you are free to scrape it. Separately, if you cannot inform people individually, it says you must at least publish the information you would normally give them, for example in a privacy policy on your website.

Special Category and Criminal Data

Special category data includes data about health, religious beliefs, political opinions, ethnic origin, and sex life, and processing it is prohibited unless an exception applies. The guidance warns that scraping, especially of social media, quickly picks up this kind of data, even unintentionally. If ordinary and special category data are stored together in a way that cannot be separated, the stricter rules apply to the whole database.

One exception covers data that the person concerned has manifestly made public. The guidance explains that the person must have published the data themselves, intending to make it accessible to a broad audience, and must have done so explicitly, through a clear, active act. A public profile under a platform's default settings does not count, because there is no active choice. The guidance applies a similar approach to criminal data. For now, the guidance says, if you cannot rule out processing these types of data without an exception, you should assume the processing is not allowed. It also notes that the CJEU may in future treat the initial collection stage of scraping the way it treats search engines, which do not have to check for such data before indexing.

What the AP Sees as More and Less Feasible

The guidance stresses that it cannot say in general terms which uses are allowed, and that every case must be assessed on its own facts. It does give examples. Scraping that is more likely to be brought in line with the GDPR includes:

Public news websites, to follow developments relevant to your own organisation or field.
Online shops scraping their own web pages, for example customer reviews, to communicate with their own customers.
Public online forums about information security, to identify security risks for your own organisation.

Examples that, according to the guidance, can probably never be made GDPR compliant include:

Scraping the internet to build profiles of people and sell them.
Scraping closed social media accounts or closed forums.
Scraping social media accounts, even public ones, to decide whether someone gets an insurance policy.

The AP has also shown it will enforce in this area. In September 2024, it announced a fine of 30.5 million euros for the facial recognition company Clearview AI over a database of photos of faces collected from the internet.

What This Means for Proxy Users

A proxy does not change the rules

Collecting data through a Dutch mobile proxy changes the IP address a website sees. It does not change which law applies or what it requires. Under the GDPR, what matters is whether personal data is processed, by whom, and, for organisations outside the EU, whether they offer goods or services to people in the EU or monitor their behaviour. The network a request comes from is not part of that test.

Pages often contain personal data too

Many proxy workflows target data that is not about people at all, such as product prices, availability, search rankings, or ad placements. The GDPR applies when personal data is processed. But pages with prices or rankings often also contain names, for example in customer reviews or search results. Because the guidance says a legal basis is needed even for data you delete right after collection, a practical approach is to decide before collecting which pages and fields you need, and to leave the rest out of your collection.

Points to consider

Based on the guidance, these are points worth considering before a project starts:

Consider defining the purpose and the exact data you need in advance, and leaving personal data out where you can.
If geography matters for scope, the guidance advises filtering the URLs you visit rather than the data after collection.
If personal data is involved, assess legitimate interest and whether a DPIA is required. The AP guidance notes that a DPIA will in many cases be mandatory and recommends carrying one out even where it is not, and that if a high risk remains, you must consult the AP before starting.
Follow internet standards such as robots.txt, which the guidance lists as a possible safeguard.
Check copyright, database rights, and website terms separately. The AP guidance does not cover them.

For examples of workflows that focus on non-personal data, see Mobile Proxies for Dutch E-commerce: Price Monitoring and Research and Tracking Google.nl Rankings with Mobile Proxies: A Dutch SEO Guide.

FAQ

Is web scraping illegal in the Netherlands?

The AP guidance does not ban scraping as such. It concerns scraping that involves personal data, and it finds that this is often difficult or impossible to do lawfully, while mainly targeted forms of scraping can be lawful. Scraping that involves no personal data falls outside the GDPR, although other rules, such as copyright, database rights, and website terms, can still apply.

Does the guidance apply to price monitoring?

Only to the extent that personal data is processed. Prices and product details are usually not personal data, but pages that contain them can also include names, for example in reviews, and seller details on marketplaces can identify individual traders.

Can a commercial interest be a legitimate interest?

It can. In the KNLTB ruling of 4 October 2024, the CJEU held that a commercial interest "could constitute a legitimate interest", provided it is not contrary to the law, and that this must be assessed case by case. The April 2025 edition of the AP guidance cites that ruling and no longer contains the exclusion of purely commercial interests. The other two conditions, necessity and the balancing test, still apply.

Does a Dutch proxy change which rules apply?

No. The IP address you use does not decide whether the GDPR applies. What matters is whether personal data is processed, by whom, and, for organisations outside the EU, whether they offer goods or services to people in the EU or monitor their behaviour.

Is the AP guidance binding law?

The guidance sets out how the AP, as the Dutch supervisory authority, interprets the GDPR. The final interpretation of the GDPR lies with the courts, as the KNLTB ruling showed.

This article summarises the AP's guidance and related sources for general information. It is not legal advice. For a specific scraping project, especially one involving personal data, consult a privacy professional.

Related Reading

● 99.9% Uptime ⚡ 4G/5G Carrier SIMs Odido, KPN, Vodafone

Dedicated Mobile Proxies in the Netherlands

PowerProxy provides dedicated mobile connections on real carrier SIMs, with HTTP, SOCKS5, and OpenVPN support. In Amsterdam, PowerProxy offers Odido, KPN, and Vodafone SIMs.

✓Odido, KPN and Vodafone SIMs in Amsterdam
✓A physical modem and SIM reserved for you alone
✓Rotation at set times or via API
✓HTTP, SOCKS5 and OpenVPN

No credit card required  ·  4.3 / 5 on Trustpilot ★★★★★  ·  1,200+ businesses

Enjoyed this article? Share it with your network
Narmin Kamilsoy
Written by

Narmin Kamilsoy

Contributing author sharing insights and stories on our blog.

WeChat Telegram