In plain words. PromptScrubber is a free utility. It helps you remove personal data before pasting text into an AI tool. It is not a guarantee, not a legal service, and not a substitute for your own review.
1. Acceptance of these terms
By using this website you agree to these Terms of Service. If you do not agree, do not use the service. The terms apply to every visitor, with or without an account — and no account exists.
2. What the service is
PromptScrubber is a static web application that runs entirely in your browser. It detects candidate personal data in text you provide, replaces it with placeholders, and can restore the original values from a mapping held in the memory of your tab.
The service is provided free of charge. No payment, subscription or registration is required, and no paid tier is currently offered. Any future paid features will be presented only when they exist.
3. Permitted use
You may use PromptScrubber for personal or professional purposes, including commercial work, provided that:
- you have the right to process the text you paste, and you comply with the data protection rules that apply to you;
- you do not use the service to commit a crime, to stalk, harass or surveil another person, or to circumvent a legal obligation;
- you do not attempt to break, overload or reverse-engineer the site in a way that harms other users;
- you do not present the service as your own product, or resell access to it as if you were the provider;
- you do not remove copyright, licence or attribution notices from the bundled third-party libraries.
4. Your responsibility
You are responsible for the text you paste and for the text you publish, send or store afterwards. Because processing happens on your device, we neither receive your document nor exercise any control over it.
You must verify the result before sharing it. Automated detection reduces risk; it does not eliminate it. Before pasting anonymized text into an AI assistant, read it once and confirm that nothing identifying remains.
5. Limits of automatic recognition
The tool combines regular expressions, checksums and local dictionaries. Some categories are reliable (email addresses, IBANs with a valid checksum, payment cards that pass the Luhn test, Italian tax codes, IP addresses). Others are inherently heuristic:
- Person and company names depend on local dictionaries and capitalisation patterns. Names outside those dictionaries, unusual spellings and single first names may be missed.
- Dates of birth are recognised only when the surrounding context makes them explicit.
- Addresses, medical details, job titles and free-form identifiers are not detected automatically. Use the manual selection feature.
- Scanned PDFs contain no text layer and are rejected: OCR is not implemented. PDF and DOCX files are converted to plain text, and the anonymized output is a
.txtfile.
No claim of complete or perfect anonymization is made anywhere on this site.
6. Third-party AI tools
PromptScrubber is independent from OpenAI, Anthropic, Google and any other AI provider. When you paste text into those tools, their terms and privacy policies apply. Anonymized text is still text — it may contain indirect identifiers such as a job title, a city and a date that, combined, re-identify someone. Judging that residual risk is your responsibility.
7. Intellectual property
The site’s interface, text and original code are protected by copyright and remain the property of the publisher identified in the Privacy Policy. The bundled libraries (PDF.js, Mammoth.js) are distributed under their own licences, included in the vendor folders.
You retain all rights to the text you process. We claim no licence over it, for the simple reason that we never receive it.
8. Availability and changes
The service is provided “as is” and “as available”. There is no service level agreement, no uptime commitment and no guarantee that a feature will remain unchanged. Features may be modified, suspended or removed, including this free version. Where a change materially affects these terms, the “last updated” date will change and the new text will apply from publication.
9. Disclaimer of warranties
To the maximum extent permitted by applicable law, the service is provided without warranties of any kind, whether express or implied, including merchantability, fitness for a particular purpose, accuracy and non-infringement. We do not warrant that detection is complete, that the output is error-free, or that the service will be uninterrupted.
10. Limitation of liability
To the maximum extent permitted by applicable law, the publisher is not liable for indirect, incidental, special, consequential or punitive damages, nor for loss of profits, data, goodwill or business opportunity, arising from the use of or inability to use the service — including any disclosure that results from relying on its output.
Nothing in these terms excludes or limits liability that cannot be excluded or limited by law, such as liability for wilful misconduct or for death or personal injury caused by negligence.
11. Privacy
Processing is described in the Privacy Policy. In short: no document content reaches us, and the only local data is a colour scheme, a detection preference and a consent choice.
12. Governing law and jurisdiction
These terms are governed by the laws of [COUNTRY]. The courts of [COUNTRY] have jurisdiction, without prejudice to mandatory consumer protections in your country of residence.
13. General
If a provision of these terms is found unenforceable, the remaining provisions stay in force. Failure to enforce a provision is not a waiver. These terms, together with the Privacy Policy, the Cookie Policy and the Disclaimer, constitute the entire agreement about the use of this website.
14. Contact
[CONTACT_EMAIL] · contact page