Question 1

What rights does AIScrapeSafe assess?

Accepted Answer

AIScrapeSafe assesses 8 rights per source: scrape, tdm (text-and-data-mining), aiTrain, aiUse, derive, redistribute, transform, and tag. Each is allowed, restricted, or unknown — and unknown is never treated as allowed.

Question 2

Which content-usage rights does AIScrapeSafe check?

Accepted Answer

AIScrapeSafe assesses 8 rights per source: scrape, tdm, aiTrain, aiUse, derive, redistribute, transform, and tag. Each is allowed, restricted, or unknown.

Question 3

What can AIScrapeSafe do?

Accepted Answer

AIScrapeSafe gives you a machine-readable usage-license verdict for any URL — whether you may scrape it, mine it (TDM), train or use AI on it, derive, redistribute, transform, or tag it — with the evidence behind each call. Paste a URL on the home page to try it.

Question 4

What is the "tag" right?

Accepted Answer

The "tag" right covers labeling or annotating a source's content (e.g. attaching metadata or classification tags). It is tracked separately from the mining/AI rights because a source can allow tagging while reserving training or redistribution.

Question 5

What is AIScrapeSafe and who is it for?

Accepted Answer

AIScrapeSafe automates data-provenance checks for AI/ML data acquisition. Point it at a URL and it returns a machine-readable verdict on whether you can use that content, with an auditable evidence trail. It's built for data-acquisition engineers (a fast API in the pipeline) and GRC/compliance analysts (a defensible record of due diligence).

Question 6

What is a Proof of Screening certificate?

Accepted Answer

It's a permanent, timestamped, auditable record of a rights query for a specific URL at a moment in time - the query details, the per-right verdict, the evidence snapshot, the logic applied, and a stable ID you can re-verify later. It turns an ephemeral check into a durable compliance asset.

Question 7

What are AIScrapeSafe's limitations at launch?

Accepted Answer

It's optimized for HTML pages with text; video, images, and JavaScript-heavy content may be limited. Legal-language analysis is most accurate in English. A verdict reflects the site's state at query time. And it is not legal advice - it's evidence of due diligence, not legal immunity.

Question 8

What are the seven signals AIScrapeSafe analyzes?

Accepted Answer

robots.txt (including AI agents like GPTBot and CCBot), Terms of Service, copyright notices, license declarations (Creative Commons, MIT, GPL), API and access terms, privacy policies, and machine-readable license standards like the OpenControls.ai schema. The set is modular - new signal types get new parsers.

Question 9

What if a site changes its terms after I get a certificate?

Accepted Answer

A certificate is a point-in-time snapshot proving you did due diligence at ingestion; it does not grant perpetual rights. For one-time corpus creation it's your record for that date. For ongoing ingestion or RAG, re-run checks periodically - changes to terms or robots.txt can be flagged.

Question 10

How do I integrate AIScrapeSafe into my pipeline?

Accepted Answer

Call the low-latency REST API (URL-keyed) directly inside your acquisition loop: check a URL, proceed if allowed, and optionally generate a Proof certificate when confidence is high. An official Python SDK is first; the API targets sub-500ms latency so it will not bottleneck.

Question 11

Can I call AIScrapeSafe from an AI agent or via MCP?

Accepted Answer

Yes - it ships a native MCP server, so agents and assistants like Claude and Claude Cowork can request a verdict inside their own workflow with no glue code, getting the same structured rights, evidence, and confidence the REST API returns.

Question 12

What's free and what do I pay for in AIScrapeSafe?

Accepted Answer

Real-time, ephemeral rights checks are free up to a generous monthly limit - enough to make in-pipeline allow/deny decisions. You pay only to generate the permanent, auditable Proof of Screening certificate, i.e. for the domains you actually ingest and need to document.

Question 13

Can I export data for an audit or regulators?

Accepted Answer

Yes. Any certificate downloads as a signed PDF (human-readable) or JSON (raw evidence), and you can bulk-export a set of checks as a CSV summary with URL, date, verdict, confidence, and a link to the full certificate.

Question 14

What data does AIScrapeSafe store, and do you train on my queries?

Accepted Answer

Free checks log only aggregated, anonymized metadata; paid certificates store the URL, verdict, and evidence linked to your account, encrypted at rest. We do not store the content you scrape or your model details, and we do not train our models on your queries - the diligence network learns only from domains' public rights signals.

Using AIScrapeSafe