Acceptable Use Policy — AI-Generated Content

Last updated: November 12, 2026

pngtext.com ("pngtext", "we", "us", or "our") provides AI-powered image generation and editing tools. This Acceptable Use Policy ("Policy") applies to all content generated, edited, uploaded, published, or stored through pngtext, and sets out our content standards, moderation process, reporting procedures, and enforcement actions. It applies to every user of the service.

This Policy is a binding part of our Terms of Service. Using pngtext means accepting it.

1. Scope

This Policy covers every kind of content that passes through the service:

• Images — pictures generated, edited, extended, upscaled, restored, background-removed, or otherwise processed on pngtext.
• User uploads — photographs, artwork, or other images you send to the service as input.
• Prompts and instructions — the text descriptions and editing instructions you submit.
• Any other content you create, store, or share through the service.

Prompts are treated as content. A prompt that describes prohibited material is a violation even if no image is ultimately produced.

2. Content Standards and Prohibited Categories

pngtext prohibits the generation, editing, or processing of the following six categories of content:

• Pornography and NSFW content.
• Violence and gore.
• Hate speech.
• Child-unsafe content, including CSAM.
• Deepfakes and impersonation of real people.
• Content that infringes copyright or trademarks.

The sections below break these categories down and add the violations that apply on top of them.

2.1 Absolutely Prohibited — Zero Tolerance

The following are actioned immediately and may be reported to law enforcement. There is no warning step and no exception:

• Child sexual abuse material (CSAM), or any sexual or sexualized depiction of a minor, whether real, illustrated, or synthetically generated.
• Content that promotes, encourages, or facilitates terrorism, violent extremism, or mass violence.
• Technical instructions for weapons of mass destruction.
• Content inciting genocide or ethnic hatred.
• Deepfakes of real individuals, including face swaps onto real people, undressing or nudification, compositing a real person into an intimate or sexual scene, and fabricating statements or actions a real person never made — whether or not the individual consents. Consent does not make this content acceptable.

2.2 Restricted — Allowed Only With Conditions

The following are not categorically banned but are subject to review and to the conditions stated:

• Violent or disturbing imagery — permitted only where there is a clear creative, journalistic, educational, or safety-related purpose, and never as gratuitous gore.
• Politically sensitive imagery — content that could be mistaken for authentic documentation of real events or real public figures is not permitted.
• Medical, legal, or financial imagery presented as advice — must comply with applicable local regulations and must not be presented as professional guidance.
• Likenesses of public figures in satirical or editorial contexts — permitted only where the result is not deceptive, is not sexual, and does not purport to be a real photograph of a real event.

Note that pornographic and NSFW content appears in Section 2 above as a prohibited category. It is not available on pngtext under any circumstances.

2.3 Generally Prohibited Conduct

In addition to the categories above, you may not:

• Generate or spread misinformation, or content presenting disproven claims as fact.
• Impersonate a real person, brand, or organization, or imply their endorsement.
• Infringe intellectual property, publicity, or privacy rights, including reproducing logos, trademarked characters, or copyrighted artwork you do not have rights to.
• Pass AI-generated content off as human-created in a context where that distinction is material.
• Bulk-generate spam or deceptive commercial content.
• Submit jailbreak prompts or any technique intended to bypass our safety measures or those of the underlying models.
• Use the service to build or operate a competing model, or to scrape outputs for model training, without written permission.

2.4 Source of Uploaded Material

pngtext accepts uploads of photographs, which may include images of real people. You may upload only:

• Your own photographs, or
• Material for which you hold explicit permission from every person depicted.

Uploading another person's photograph without their consent is prohibited. Before uploading, you confirm that the material is yours or that you have permission. Holding permission to use a photograph does not authorize generating the deepfake content prohibited under Section 2.1 — permission to use an image is not permission to fabricate what that person said, did, or looked like.

If you are uploading an image of a minor, you must be the parent or legal guardian, or hold documented permission from them.

3. Content Moderation

pngtext operates a multi-layer moderation system combining automated review, human review, and periodic auditing. No request reaches the image models without passing an automated check, and no raw-model passthrough is offered.

3.1 Automated Review (Real Time)

Every prompt and every uploaded image is screened before generation, and every generated result is screened after.

• Input filtering — prompts are checked for prohibited keywords, semantics, and intent.
• Image input filtering — uploaded images are screened for prohibited material before processing.
• Output filtering — generated images are assessed for safety before being returned.
• Usage pattern analysis — accounts are monitored for anomalous or automated abuse.
• Moderation provider — we use the Waffo Prompt Sift moderation API, together with the safety filters built into the underlying image models, to screen prompts and generated output.
• Blocking behaviour — a request that fails moderation is rejected, no image is produced, and no credit is consumed for the blocked request.

3.2 Human Review

• Content escalated by the automated layer enters a human review queue.
• Content reported by users is reviewed manually by our trust and safety team.
• Reports concerning child safety are escalated to the top of the queue and reviewed on a best-effort basis around the clock.
• Reviewers are bound by confidentiality obligations.

3.3 Periodic Audits

• We sample generated content at random on a recurring basis to evaluate filter performance.
• Moderation classifiers are reviewed and adjusted when new evasion patterns appear.
• Accounts with a history of violations are audited more frequently.

3.4 Content Severity Levels

• L1 — Critical. Child safety, terrorism, WMD instructions. Auto-blocked, confirmed by a human, immediate permanent ban, reported to the relevant authority where required.
• L2 — High. Severe violations, including non-consensual sexual deepfakes and hate content. Auto-flagged, priority review, content removed, account suspended.
• L3 — Medium. General violations, such as copyright infringement or commercial deception. Human review, content removed, warning issued.
• L4 — Low. Minor or first-time violations with no clear harm. Report-triggered, warning issued and edit or removal required.

4. Reporting

4.1 When to Report

Please report content or behaviour that:

• Violates the content standards in Section 2.
• Involves child safety, terrorism, or non-consensual intimate imagery — these are urgent.
• Impersonates you or infringes your own rights, including your likeness, copyright, or trademark.
• Otherwise appears to misuse the service.

4.2 Reporting Principles

• Confidentiality. The identity of a reporter is protected and is not disclosed to the reported party.
• Impartiality. Every report is reviewed on its merits, whether it comes from a user, a rights holder, or a third party.
• Anti-abuse. Knowingly false or malicious reports are recorded and may result in action against the reporting account.

5. Reporting Channels

Report any content or behaviour that violates this Policy by emailing support@pngtext.com. This one address receives content reports, copyright and trademark complaints, privacy and data requests, and general support; it is monitored and triaged on arrival.

When reporting, please include the image or a link to it, a description of the problem, and — for rights complaints — a statement that you hold the rights or are authorized to act for the rights holder. We may ask for additional information to verify a report.

6. Response and Processing Timelines

• L1 — Critical: initial response within 2 hours, resolution within 24 hours.
• L2 — High: initial response within 24 hours, resolution within 3 business days.
• L3 — Medium: initial response within 3 business days, resolution within 7 business days.
• L4 — Low: initial response within 5 business days, resolution within 15 business days.

Where a case needs longer, we contact the reporter with the reason and a revised estimate.

Processing runs in this order: report received and acknowledged automatically, triage and severity classification, review by automated and human reviewers, decision, enforcement action, notification to the reporter and the affected account.

7. Enforcement Actions

We apply the least severe action that resolves the issue, escalating on repeat behaviour:

1. Warning — written notice requiring removal or correction. Applied to first-time minor violations.
2. Content removal — the violating content is taken down. Applied to confirmed L3 and L4 cases.
3. Feature restriction — reduced generation quota or restricted model access. Applied to repeat minor violations or a single L3 case.
4. Account suspension — temporary suspension pending review. Applied to L2 cases or cumulative violations.
5. Permanent ban — permanent termination of access, with any remaining credit forfeited. Applied to L1 cases or repeat L2 cases.
6. Referral to authorities — for criminal-level L1 content, we report to the relevant law enforcement body and cooperate with any investigation.

We may also preserve and disclose content and account records where legally required, or where necessary to investigate a violation.

8. Appeals

If you believe an enforcement decision was wrong, you may appeal within 30 days of receiving notice:

• Appeal channel: support@pngtext.com
• What to include: your account email, the content or action concerned, and the reason you believe the decision was incorrect.
• Timeline: appeals are reviewed within 10 business days.
• Review method: appeals are decided by a reviewer who was not involved in the original decision.

Appeals do not apply to confirmed CSAM findings, which are never reinstated.

9. Transparency and Policy Updates

• We review this Policy at least once a year and whenever the product's capabilities or the applicable rules change.
• Material changes are announced on this page at least 15 days before taking effect, and the date at the top is updated.
• Continued use of pngtext after a change takes effect constitutes acceptance of the revised Policy.
• We publish aggregate data on reports received and actions taken on request from a regulator or a payment partner.

10. Contact Us

• Email: support@pngtext.com
• Handles: content reports, appeals, legal and rights complaints, privacy requests, and general support.
• Operator: pngtext.com
• Website: https://pngtext.com

This Policy sets out our content standards and enforcement practice. It does not create rights for any third party, and it does not replace the Terms of Service.