Terms and Conditions

The agreement

zeromark.ai (below: zeromark, we, us) provides this website and the tools described on it. By using the site or any of those tools you accept these terms. If you do not accept them, please do not use the service.

The service is operated by [to be confirmed]. If you use it on behalf of an organisation, you confirm that you are allowed to bind that organisation to these terms.

You need to be old enough to enter into a contract where you live. If you are not, use the service only with the consent of a parent or guardian.

What the service is today

Right now the service is this website, the free cleaner on the home page, and the browser extension. The cleaner removes hidden and unwanted characters from text you paste into it: zero-width characters, Unicode tag payloads, bidirectional controls, leftover assistant chatter, and formatting artefacts.

The cleaner is a deterministic program that applies fixed rules. It contains no language model, performs no inference, and makes no network request with your text. All of it happens in the browser tab you have open.

It does not rewrite your sentences and does not swap your word choices. Rule groups you switch off are not applied.

The browser extension is published for Firefox. Its Chrome listing, the mobile share action, and the governance API described elsewhere on this site are in development and not yet available. Nothing here is a promise that they will ship, or that they will behave as described when they do.

Responsible use of AI generated text

This tool exists because language models leave marks in text that have nothing to do with what the writer meant. Removing that clutter from your own writing is legitimate. Using the same capability to deceive someone who is entitled to the truth is not, and it is not what we built it for.

We ask you to follow four principles, drawn from widely accepted guidance on responsible AI, including the OECD AI Principles and the transparency duties in Article 50 of the EU AI Act, which have applied since 2 August 2026:

• Honesty about authorship. Cleaning formatting is not the same as denying that a model was involved. When a person, a platform, an employer, a client, a school, or the law asks whether content is AI generated, answer truthfully.
• Human oversight. Read what you publish. A model can be confidently wrong, and removing the marks that show a machine wrote something does not make its claims correct.
• Accountability. Whoever presses publish owns the result. That is you, not the model and not us.
• Respect for other people. Do not use cleaned text to impersonate, to harass, to mislead anyone about health, money, or products, or to break a law that applies to you.

Some disclosure duties are legal duties, not preferences. In the European Union, providers of generative systems must mark synthetic content so that it can be detected as artificially generated, and deployers must label certain material, including deepfakes and text published to inform the public on matters of public interest. Removing a marking in order to defeat such a duty, or to help someone else defeat it, is a misuse of this tool and a breach of these terms. Where a disclosure duty applies to you, meet it in your own words, even when no marker survives in the text.

Things you must not do

In particular, you must not use the service to:

• present machine generated work as your own where the rules that bind you require disclosure, including in academic and professional settings
• strip provenance or watermark information in order to evade a legal, contractual, or regulatory disclosure duty
• produce or launder content that defrauds, impersonates, defames, or harasses
• process text you have no right to process, including other people's confidential information or personal data
• resell, sublicense, rebrand, or decompile our engine, or present its output as your own product
• attack the service, by seeking unauthorised access, overloading it, scraping it at a scale that degrades it, or circumventing security controls
• break any law that applies to you, or help anyone else do so

We may refuse or withdraw access if we believe the service is being used in these ways. We should be straightforward about the limits of that: the free cleaner runs on your own device, so this is a matter of trust rather than something we can technically police. What we can do is state plainly what the tool is for.

You remain the author

You are responsible for the text you put into the tool and for everything you do with what comes out. That includes checking facts, quotes, and links, and checking that publishing it is lawful where you are.

We do not review, approve, or endorse anything processed by the tool. Cleaning is a formatting operation. It is not a quality certificate and not a compliance sign-off.

Nothing on this site is legal, regulatory, financial, or other professional advice. If you need to know whether a disclosure duty applies to you, ask a qualified adviser.

What the tool does not claim to do

We want to be exact here, because the opposite claim is common and wrong.

• The tool does not make text undetectable as AI generated, and we do not offer it for that purpose. Detectors look at style and statistics, not only at hidden characters.
• Removing a marker does not remove the fact. If a model generated the content, a model generated it, whatever the bytes say.
• The tool cannot find every marker. Marking techniques change, and new ones will appear that our rules do not cover.
• Of the four guarantees described on this site, only the removal of hidden markers and machine clutter ships today. Grounding checks, link verification, and prompt injection detection are roadmap items and are labelled as such.

Your text and your rights in it

Your text stays yours. You grant us no licence to it, because we never receive it. The cleaner processes it in your browser and keeps no copy. Close or reload the tab and nothing remains.

You are responsible for having the right to process whatever you paste, and for not pasting material you are not allowed to handle, such as other people's personal data or anything covered by a confidentiality obligation.

Privacy and data

Text you clean is never transmitted to us. There is no upload, no server-side processing, and no log of it. That is a design constraint of the engine, not only a policy.

The one place we collect personal data is the early access list. If you submit that form we store the email address you give us, the tier you said you were interested in, an optional company name, the site language you used, and the time of the request. The record lives in a database hosted in the European Union. We do not store IP addresses or user agents, and we do not log submissions.

We use that address for one purpose: to write to you once, when the tier you picked becomes available. No newsletter, no advertising, and no sharing or selling to anyone. Every message carries a one-click unsubscribe link, and you can ask us to delete your record at any time.

The site sets no advertising or analytics cookies and runs no tracking scripts. Your browser stores your language choice locally so the site can remember it. Web fonts are loaded from Google Fonts, which means your browser contacts Google servers and Google receives your IP address as part of that request. The site itself is served from Cloudflare, which processes requests as our hosting provider.

Our intellectual property

The zeromark name, the wordmark, the site design, the text on it, and the sanitizer engine belong to us and are protected by copyright and trade mark law. These terms let you use the tool as offered and nothing more. They transfer no ownership and grant no licence to copy or redistribute the engine.

Third party open source components keep their own licences. This site is built on a template released under the MIT licence, and the relevant notices are kept in the source.

Plans, prices, and availability

The web cleaner is free. It needs no account and no payment, and we intend to keep it that way.

The paid tiers on the pricing page are not on sale. There is no checkout, we operate no payment processor, and joining the early access list creates no contract, no obligation to buy, and no charge. Prices shown are indicative and may change before launch.

When paid plans do open they will have their own terms covering billing, renewal, refunds, and service levels, and those terms will be shown before any payment is taken.

Development status and changes

This is an early product and parts of it are unfinished. Features may change, be delayed, or be dropped, and we may modify or discontinue any part of the service. One thing we will not do quietly is start sending your text somewhere: if that ever changed, it would be a different product and it would be announced.

The site may contain errors, and translated pages can lag behind the English original. Where a translation and the English version conflict, the English version governs.

Third party services

We rely on providers to run the service: currently Cloudflare for hosting, static assets, the early access database, and bot protection, and Google Fonts for typefaces. Their handling of data is governed by their own terms. Future paid products may route requests to model providers, including AWS Bedrock, and that will be documented before those products are offered.

Links to other sites are offered for convenience. We do not control them and are not responsible for their content.

No warranty

The service is provided as it is and as it becomes available, without warranties of any kind, express or implied, including any implied warranty of merchantability, fitness for a particular purpose, accuracy, or non-infringement.

We do not warrant that the tool will catch every marker, that it will be free of defects, that it will be available without interruption, or that its results will meet any particular legal or editorial standard.

Some jurisdictions do not allow the exclusion of certain warranties. Where that is so, the exclusions above apply only as far as the law permits, and your statutory rights as a consumer are unaffected.

Limitation of liability

So far as the law allows, we are not liable for indirect, incidental, special, consequential, or punitive damages, nor for lost profits, lost data, reputational harm, or regulatory penalties arising from your use of the service.

Because the current service is provided free of charge, our total liability for any claim relating to it is limited to what you have paid us for it, which for the free tool is nothing. Where a paid plan exists, its own terms will state the applicable cap.

Nothing in these terms excludes liability that cannot be excluded by law, such as liability for death or personal injury caused by negligence, or for fraud.

Suspension and termination

You can stop using the service whenever you like. There is no account to close for the free tool. If you are on the early access list you can remove yourself with the unsubscribe link, or by writing to us.

We may suspend or withdraw access, wholly or partly, if we believe these terms have been broken or if we need to protect the service, its users, or ourselves. The clauses on your responsibilities, intellectual property, warranties, and liability continue to apply after you stop using the service.

Changes to these terms

We may update these terms when the product changes or when the law does. The date at the top of this page shows the last revision. If a change is significant we will make that clear on the site. Continuing to use the service after a change means you accept the revised terms, and if you do not accept them, please stop using it.

Governing law

These terms, and any dispute about them or about the service, are governed by the law of [to be confirmed], and the courts of [to be confirmed] have jurisdiction. If you are a consumer, this does not remove the protection of the mandatory consumer law of the country where you live, or your right to bring proceedings there.

We would much rather resolve a problem by email than in a courtroom. Please write to us first.

Contact

Questions about these terms, requests to delete your data, and reports of misuse all go to hello@zeromark.ai.