What to worry about, in order, and the profile to hand on

1What to be concerned about, in order

Every concern the checkup can raise, from the cheapest to deal with to the one that needs a lawyer's view, each linking the document that explains it. With the script on, only the ones your answers raise.

1. terms of use

   Read the site's terms page before you fetch anything. If it says no to automated access, or to the use you plan, that is the cheapest problem to find and the most expensive to ignore. Read the document.

2. robots.txt and the other crawl signals

   Read the site's robots.txt and look for a reservation: a “no AI training” line, a TDM header, a licence. In some places it binds you; in most it is evidence against you. Read the document.

3. terms of use

   The pages are behind a login. You will be inside a contract, and the public-page rule that helps you under computer-misuse law no longer does. Read the document.

4. where the site is

   You could not say where one or more of the target sites is. Work it out with the evidence ladder, and until you can, assume the stricter of the places it could be and write down that you could not tell. Read the document.

5. the people in the data

   The pages have people on them. You need a lawful basis, a notice people can find, and a way for them to say no. “It was public” is not the basis. Read the document.

6. computer misuse and unauthorised access

   Something is in the way. Getting past a technical barrier is where computer-misuse law starts, and in July 2026 a US court treated a challenge in front of a page as a lock under anti-circumvention law. Read the document.

7. computer misuse and unauthorised access

   You have been told to stop, or blocked. Continuing after a letter or a block breaks the site's terms everywhere, and in many places it turns the next request into unauthorised access; on a public page in the US, the courts are not agreed. Read the document.

8. copyright and text-and-data mining

   You will publish, sell or train on what you collect. Copying is the act copyright reaches. Whether an exception covers you depends on the place and the use, and in the EU a reservation switches the commercial exception off. Read the document.

9. database rights

   In the EU and a few other places a compiled database is protected on its own, whatever the copyright in its contents. Taking a substantial part of one is a claim by itself. Read the document.

10. unfair competition

    You are building something that competes with the site using its own compiled work. That is the fact pattern behind hot-news and unfair-competition claims. Read the document.

11. AI law

    A model is in the loop. Its provider's place adds that place's law, and a model trained on what you collect brings the AI Act's duty to honour reservations. Read the document.

12. the people in the data

    The data will leave the country it was collected in. Transfer rules apply on the way out. Read the document.

13. who runs it for you

    A vendor scrapes for you. Their place and their conduct are yours to answer for in the places you are. Put every question above to them, in writing. Read the document.

2The places

With the script on, each place your answers name appears here, linked to its law page where LexLint has written one. Every place is on the law index.

3The profile, for a developer or a lawyer