The dataset is free.
The evidence record stays complete.

The evidence is complete on every plan: origin, licences, preparation operations, quality checks; bias assessment not performed, and personal-data findings. The preview status is inside the body its SHA-256 covers, so changing it breaks the digest. Current status: complete preview; independent issuance is in verification rollout.

No, and any vendor who tells you otherwise is selling you a risk. DATAtokeniz generates the data-governance documentation Article 10 asks providers to keep, derived from what the pipeline actually observed. It is evidence for a compliance file. It is not legal advice and it is not a conformity assessment — you still need those.

Several, and they are routinely conflated — the Digital Omnibus adopted in June 2026 moved some of them. Obligations for general-purpose AI models have applied since 2 August 2025; models placed on the market before that date have until 2 August 2027 to come into full compliance. On 2 August 2026 two separate things happen: Article 50 transparency duties begin, and the Commission and national authorities gain the power to impose fines — so that date is an enforcement date, not the start of the obligation. Watermarking under Article 50(2) has a grace period to 2 December 2026 for systems already on the market. Article 10 data governance applies to Annex III high-risk systems from 2 December 2027 and to Annex I from 2 August 2028. DATAtokeniz labels every record as directly extracted or model-generated — lineage evidence that supports Article 50 work, though the disclosure and machine-readable marking of generated content in your own product remains your obligation — and compiles the Article 10 record for the later dates. The Digital Omnibus (Regulation (EU) 2026/1744, in force 27 July 2026) also added Article 4a, carrying the bias-detection data basis beyond high-risk systems. Which articles bind you depends on your role and risk class — confirm that with counsel, not with us.

Once per source page, not once per output row. A 40-page PDF that yields 300 records counts as 40 pages. Counts are measured from the finished dataset, so any invoice line can be reconstructed from the dataset itself.

Eligible new Pro customers can receive a 14-day free trial, once per Stripe customer. A card is required up front so the subscription can continue without interruption. Stripe Checkout shows your eligibility and exact charge date before you confirm. If Stripe says you are not eligible, no trial is applied and the charge follows the date shown there. Starter is sold without a trial — at $29 the free tier is already the trial.

Volume. Starter raises the ceiling — 2,000 pages and 100 runs a month against Free's 500 and 20 — and Pro raises it to 5,000 pages and 500 runs. Datasets, per-row provenance, quality reports and complete Article 10 preview findings remain available. Attestation status: complete preview; independent issuance is in verification rollout.

The run is refused until the next period, on every plan — DATAtokeniz never bills you for anything you did not explicitly buy, so there is no overage charge and no surprise invoice. If a limit is the wrong shape for your work, moving up a tier takes one click in the customer portal and prorates automatically.

From the customer portal, linked in your account menu. Upgrades, downgrades, switching between monthly and yearly, and cancellation are all self-serve and take effect without contacting us. Cancelling stops the next renewal; you keep the plan until the period you already paid for ends.

Nowhere DATAtokeniz owns. Inputs are processed in the request and never written to DATAtokeniz storage; results go to your own Google Drive. Drive access is drive.file scope only — files you pick or that DATAtokeniz creates.

On every tier the SHA-256 digest is recomputable from the record body, which detects any edit to the findings. Independent proof status: complete preview; independent issuance is in verification rollout.