AIScrapeSafe
Usage-rights oracle for the open web

Know what a site or document actually lets you do with AI.

One machine-readable verdict on scraping, text-and-data-mining, and AI-training rights. All grounded in each site's or document's own signals, with the evidence shown. Not legal advice; the real thing it reflects.

  • Evidence-backed
  • Registered & verifiable
  • API & MCP ready
780sites
7,326documents
assessed & counting

Optional — your jurisdiction and intended use (where you'll use the data, not the site's). They're recorded on the verdict and tune the legal disclaimer.

Three rights, never guessed

Color is never the whole story

Every right pairs a state with an icon and a word — allowed, restricted, or honestly unknown. We never dress up “we don't know” as a yes.

Scraping & crawling

Reads robots.txt, terms, and access signals to tell you whether automated collection is permitted, restricted, or unstated.

Text & data mining

Surfaces TDM reservations and opt-outs — including EU Article 4 machine-readable rights reservations — so you mine within bounds.

AI training & use

Separates “train a model on this” from “use this at inference” — the two rights sites increasingly treat very differently.

From URL to defensible verdict

A clean three-step path — and every verdict carries its evidence and a resolvable license ID.

1

Drop in a domain

Paste any URL. Optionally add your jurisdiction and intended use to tune the result.

2

We read the signals

robots, terms, TDM reservations, licensing channels — assembled against our rules engine.

3

Get a license you can cite

A registered, verifiable verdict with the evidence trail behind every right.

Check your first site free

See the verdict, the rights grid, and the evidence in seconds.

Analyze a site