Acceptable Use Policy

^{Last updated: 06/05/2026}

This Acceptable Use Policy (“AUP”) supplements the Terms of Service and applies to everyone who uses Diglot — the website at diglot.ai, the web editor, browser extensions, desktop and mobile apps, and the Authorship Certificate feature (collectively, the “Service”). The Service is operated by Diglot OÜ, registered in the Republic of Estonia.

This AUP exists for two reasons:

To keep the Service safe, lawful, and useful for everyone — especially ESL writers and students, our primary audience.
To keep our access to the third-party AI providers we depend on (OpenAI, Anthropic, Google, and others). Their abuse-detection systems see traffic at the level of our API keys, not individual users — so a single bad actor on our platform can cost the Service for everyone.

Breaking this AUP can lead to immediate suspension or termination of your account, with no refund, and in some cases reporting to law enforcement.

1. Things you must never do

1.1 Child safety — zero tolerance

You must not use the Service to:

create, request, transmit, store, or solicit child sexual abuse material (CSAM), real or AI-generated;
groom, sexually solicit, or facilitate the exploitation of any minor;
promote eating disorders, self-harm, or dangerous “challenges” to minors.

We treat any such activity as a hard violation. We will permanently ban the account and, where applicable, report the activity (with retained logs) to the U.S. National Center for Missing & Exploited Children (NCMEC) or the relevant authority in your jurisdiction, without prior notice.

1.2 Weapons, mass harm, and dangerous capabilities

You must not use the Service to design, plan, or facilitate:

biological, chemical, radiological, or nuclear (“CBRN”) weapons;
explosives, mass-casualty attacks, or violent extremism;
malware, ransomware, or cyber-weapons targeting infrastructure, hospitals, or schools.

1.3 Illegal activity, fraud, and dangerous “high-stakes” automation

Don’t use the Service to commit, plan, or facilitate any criminal activity, including fraud, financial crime, money laundering, identity theft, harassment, stalking, or hacking.

The Service is a writing assistant. Don’t use its output as the sole basis for high-stakes decisions about real people, including:

medical, psychological, or pharmacological advice;
legal advice, litigation strategy, or court filings;
decisions on hiring, firing, promotion, immigration, or insurance;
final academic grading, disciplinary outcomes, or expulsion;
creditworthiness or financial-services decisions;
emergency response.

If you use the Service in any of these contexts, a qualified human must review and own the final decision.

1.4 Privacy, biometrics, and surveillance

You must not use the Service to:

profile, deanonymise, or build dossiers on identifiable individuals without a clear lawful basis;
generate non-consensual intimate imagery, “deepfakes” of real people, or content meant to impersonate or defame;
conduct facial recognition, social scoring, or covert tracking;
upload personal data of third parties (classmates, teachers, patients, employees) that you do not have permission to process.

1.5 Manipulation, disinformation, and harassment

Don’t use the Service to:

generate spam, mass-marketing junk, link-farm content, or SEO sludge;
run political-campaign operations, election-interference content, or large-scale astroturfing;
create content for phishing, scam, or social-engineering campaigns;
produce hate speech, harassment, or content that incites violence against a person or group.

1.6 Intellectual-property infringement

Don’t use the Service to:

reproduce material protected by copyright at scale (e.g., wholesale rewrites of paywalled articles or books) without permission;
circumvent technical protection measures (DRM, paywalls);
impersonate another author, publisher, or trademark holder.

You are responsible for confirming you have the rights to whatever you upload as input.

1.7 Sanctions and export control

You must not use the Service if you are subject to a comprehensive U.S., EU, UK, or UN sanctions program (including being on an SDN, Consolidated, or equivalent list), or if you are accessing the Service from a country subject to a comprehensive embargo (currently Cuba, Iran, North Korea, Syria, the Crimea / Donetsk / Luhansk / Zaporizhzhia / Kherson regions of Ukraine, and any region added by future law).

You also agree not to re-export or transfer the Service in violation of U.S. (EAR/OFAC), EU, or UK export-control laws.

2. Academic integrity

Diglot is a writing assistant for ESL students and professionals. It is not a service for buying essays, ghostwriting graded coursework, or evading academic-integrity systems. You agree:

Follow your school’s rules. Whether AI assistance is allowed depends on the assignment, the course, and the institution. The default in higher education is disclose, then decide — when in doubt, ask your instructor.
Do not pass off AI-generated drafts as your own unedited work in graded contexts where AI is forbidden or undisclosed. Don’t use the Service to ghostwrite an “assessment task” you are required to complete personally (for example, exam responses, capstone work, or accredited submissions under TEQSA / AAIN, FERPA-covered work, or comparable rules elsewhere).
Disclose AI use when your school, journal, employer, or course requires it.
Authorship Certificate is not a get-out-of-jail card. As stated in the Terms of Service §4, the Certificate is a record of process telemetry inside the Diglot editor — not a legal warranty of originality and not a guarantee that any institution will accept it. The burden of academic integrity stays with you.

If a school or accreditation body finds that you used Diglot in a way that violated their academic-integrity policy, that is between you and them. We cannot intervene in your case, and we are not a party to any disciplinary or honour-code proceeding.

3. Privacy of the texts you upload

You agree that you will not upload to the Service:

protected health information (“PHI”) covered by HIPAA, the EU Health Data Space, or comparable laws — unless you have your own legal basis and a separate written agreement with us;
third-party FERPA-covered education records, employee records, or other personal data of identifiable people, without their consent or another lawful basis;
payment-card data (PAN/CVV/track data), bank credentials, or unencrypted authentication secrets;
classified, export-controlled, or trade-secret material that you are not authorised to share with a U.S./EU SaaS vendor and its AI sub-processors.

If you accidentally upload such data, contact legal@diglot.ai so we can purge it.

4. Technical use — stay inside the rails

4.1 Rate limits and fair use

Plans on Diglot include monthly and daily limits. Above those limits, the Service is not “fair use.” In particular, you must not:

run automated scripts, bots, scrapers, or headless browsers against the Service except via our official API and within the rate limits;
share a single account between many people (account-sharing) — institutional or team use must use a B2B / education plan;
create multiple accounts to evade quotas or trial limits;
attempt to flood, throttle, or otherwise degrade the Service for other users.

We may impose additional dynamic limits on individual accounts that show abusive patterns, with or without notice.

4.2 No reverse engineering, no scraping the model

Except where mandatory law gives you a non-waivable right to do so, you must not:

decompile, disassemble, or reverse-engineer the Service or any of our models or prompts;
run prompt-injection, jailbreak, model-extraction, or red-team attacks on the Service outside an authorised security-research arrangement (write to security@diglot.ai if you want one);
probe, port-scan, or stress-test the Service, except inside a Diglot bug-bounty programme;
copy or republish at scale the prompts, system messages, or other internal artifacts of the Service.

4.3 No using Diglot output to build a competing product

You must not use outputs of the Service to:

train, fine-tune, evaluate, or distill another large language model, text generator, paraphraser, AI detector, or competing writing assistant;
build a public benchmark of Diglot outputs;
mass-publish raw Diglot outputs as if they were independent reference content.

This includes outputs from our free tier.

4.4 Mirroring of upstream AI providers

Because we run on third-party AI models, your use of the Service is also bound by the most recent acceptable-use rules of those providers, including:

OpenAI Usage Policies,
Anthropic Usage Policy,
Google Generative AI Prohibited Use Policy,
and any other provider we add to our Subprocessors page.

If those policies are stricter than this AUP, the stricter rule wins.

5. Transparency you owe to others

In line with the EU AI Act, when you publish or share content created with the Service in contexts where readers might reasonably assume it was written by a person:

Disclose AI use if it is required by law (e.g., political advertising, deepfakes), by the platform you publish on, by your school or employer, or by your journal.
Do not strip our AI-content notices from outputs that include them.
Take responsibility for the final text. You — not Diglot — are the publisher, and you must verify accuracy before relying on output.

6. Enforcement

We may, with or without prior notice and at our sole discretion:

Remove or block specific content;
Throttle or rate-limit an account;
Suspend or terminate your account;
Disable specific features (for example, the Cowriter agent) on your account;
Cancel your subscription with no refund for the unused portion if your conduct triggered the action;
Cooperate with law enforcement, NCMEC, or other authorities by sharing relevant data, including logs, prompts, and metadata, where we are legally permitted or required to do so;
Pursue legal remedies, including injunctive relief and damages, for violations that harm us or our users.

We try to apply enforcement proportionally. For ambiguous or borderline cases — false positives in our automated abuse-detection, for example — you can appeal to support@diglot.ai and a human will review your case. For Section 1.1 (CSAM) and Section 1.2 (CBRN / mass harm) violations there is no appeal: we ban and report.

7. Reporting abuse

If you see something on Diglot that violates this AUP — content created with our tools and used to harass, defraud, or harm someone, suspected exploitation of minors, IP theft, or anything else that worries you — please report it:

Abuse and trust & safety: abuse@diglot.ai
Security vulnerabilities: security@diglot.ai
Privacy / data-rights requests: legal@diglot.ai
General support: support@diglot.ai

Include enough information for us to reproduce or locate the content (account email if known, URL, document ID, timestamp). We will acknowledge reports within 5 business days.

8. Changes

We may update this AUP as the Service, the law, or our upstream providers’ rules change. Material changes are communicated by email or in-product notice. The “Last updated” date at the top of this page always reflects the current version. Continued use of the Service after the effective date is your acceptance.

Diglot OÜ — Republic of Estonia. Questions: legal@diglot.ai.

Resources