CC Corpus Claims

Forensic evidence bureau · AI training-data exposure

When your licensed content ends up in an AI training corpus, you need evidence, not speculation.

Corpus Claims produces exhibit-grade forensic reports establishing whether — and exactly how — your paywalled or licensed material was captured by web crawlers and carried into the datasets that train large language models. Built for publisher general counsel and plaintiff-side litigation teams.

Run a free exposure check See our services

Independent · confidential · methodology documented for discovery

What we do

Three forensic services, one chain of custody

01

Corpus detection

We trace your domain and specific works through Common Crawl and the derived training datasets built on it — FineWeb, Dolma, C4, RefinedWeb and others — documenting which URLs were captured, when, and in what volume.

Deliverable: exhibit-grade detection report, per matter.

02

Removal verification

Sent a removal demand? We independently verify whether content was actually purged — or merely reported as removed. Ongoing monitoring catches re-ingestion and silent non-compliance across subsequent crawls.

Deliverable: monitoring retainer with dated attestations.

03

Crawler attribution

We reconstruct which crawlers accessed your content, correlate capture timestamps against your access logs and terms of use, and attribute exposure to specific collection events for use in pleadings.

Deliverable: attribution forensics, expert-ready.

Why now

The litigation wave is already here

Publishers are moving from complaints to discovery, and the factual record around training-data provenance is being built right now.

  • Publishers are suing. The New York Times, CNN, McClatchy, Ziff Davis and others have active copyright litigation against AI companies over the use of their content.
  • Removal demands have gone out. Publishers have formally demanded that Common Crawl remove their material from its archive.
  • Compliance was faked. Common Crawl was reported to have represented content as removed while it remained accessible — the exact gap independent verification is built to close.
  • Provenance is contested. Whether a specific work sat in a specific corpus on a specific date is a question of evidence. That is what we produce.

Public reporting summarized for context; not a characterization of any party's legal position.

Live demo

Corpus Exposure Check

Enter a domain to query the public Common Crawl index against its three most recent snapshots. This is a fast, surface-level preview of the kind of capture data a full forensic engagement examines in depth.

Enter a bare domain (no https://, no path). The Common Crawl index can be slow; a check may take several seconds.

How it works

From intake to admissible exhibit

  1. 1

    Scope & intake

    You identify the works, domains and time windows at issue. We scope the corpora and crawls to examine and agree on evidentiary standards up front.

  2. 2

    Forensic collection

    We query crawl indices and derived datasets with documented, reproducible methodology, preserving raw records under chain of custody.

  3. 3

    Analysis & attribution

    Captures are correlated against your access logs, terms and removal timeline to establish what was taken, when, and by which collection event.

  4. 4

    Exhibit-grade report

    You receive a report written for the record — declarations, appendices and reproducible queries your experts and counsel can stand behind.

Engagements

Pricing

Detection report

$25K – $75K/ matter

A complete forensic detection engagement for a defined set of works and domains.

  • Corpus & derived-dataset tracing
  • Per-URL capture timeline
  • Documented, reproducible methodology
  • Exhibit-grade written report
Retainer

Removal monitoring

$3K – $8K/ month

Ongoing independent verification that removal demands are honored across subsequent crawls.

  • Continuous re-ingestion monitoring
  • Dated compliance attestations
  • Alerting on silent non-compliance
  • Quarterly summary for counsel

Crawler attribution

Custom/ scope

Deep attribution forensics correlating captures to specific crawlers and collection events.

  • Access-log correlation
  • Collection-event reconstruction
  • Expert-ready findings
  • Deposition & testimony support

Have a matter that may involve corpus exposure?

Tell us the publisher, the works at issue and where you are in the process. We reply to counsel and rights holders directly.

Contact Corpus Claims

or email ebo@testmachine.ai