AI Acceptable Use Policy

Version 2026-08-21 — effective as of September 1, 2026

Trakto puts generative artificial intelligence in the hands of people who create: generation of images, videos, voices, text, layouts and animations. With that power comes responsibility. This Policy defines what you may and may not do with the Platform's AI features, how we moderate, and what happens in case of a violation.

It is part of the Terms of Use and applies to all use of AI on Trakto — through the editor, Flow, the API, the MCP server or any other channel.

1. Principles

1.1. You are the author. AI is a tool. Everything you generate, publish or distribute with Trakto is your responsibility, as if you had produced it by hand.

1.2. Real people deserve protection. Faces, voices, names and works of real people may only be used with the right or consent to do so.

1.3. Transparency. Synthetic content must not deceive. When the law, the destination platform or the context requires it, disclose that the content was generated or altered by AI.

1.4. Your data does not train models. Prompts, uploads and outputs are not used to train AI models, neither by Trakto nor by the providers (details in the Privacy Policy).

2. How AI works on Trakto

2.1. Trakto combines its own logic with models from third-party providers accessed via API (Google Cloud Vertex AI / Gemini, fal.ai, OpenAI, ElevenLabs, Cutout.pro, among others listed in the Privacy Policy). Each provider has its own usage policies, which apply in addition to this Policy. Where a provider is more restrictive for a given feature, the stricter restriction applies.

2.2. Providers are used under commercial no-retention and no-training arrangements. They may, however, apply their own safety filters that block certain prompts; in such cases, Trakto will inform you that the generation was refused, with no credits consumed yes: a generation blocked by a provider filter consumes no credits in any flow.

3. Prohibited uses

It is prohibited to use Trakto's AI features to create, edit, request or distribute:

3.1. Deepfakes and non-consensual depictions

• Images, videos or audio depicting a real, identifiable person saying or doing something they did not say or do, without that person's express consent.
• Cloning or imitating the voice of a real person without their documented consent.
• Use of the face, voice, name or other identity attributes of public or private individuals in contexts that could mislead third parties about authorship, endorsement or participation.

3.2. Sexual content and nudity

• Explicit nudity, sexual or pornographic content of any kind.
• Sexualized content involving persons under 18, real or synthetic, including in illustrated or "aged-up" style. This violation results in immediate account termination and a report to the competent authorities, as required by law.
• "Undressing" (nudification) of images of real people, with or without consent.

3.3. Violence, hate and harassment

• Glorification, incitement or instruction of violence, terrorism, self-harm or suicide.
• Content that promotes hatred, discrimination or dehumanization based on race, ethnicity, origin, religion, gender, sexual orientation, gender identity, disability, age or any protected characteristic.
• Harassment, intimidation, threats or humiliation directed at identifiable persons.
• Gratuitous graphic depiction of gore or cruelty.

3.4. Fraud, deception and manipulation

• Scams, phishing, fake websites or ads, forgery of documents, seals, certificates or identities.
• False attribution: content that pretends to come from a news outlet, government body, company or person that did not produce it.
• Fabricated reviews, testimonials or social proof presented as genuine.
• Content designed to manipulate financial markets or mislead consumers about products and services.

3.5. Misleading political and civic content

• Disinformation about electoral processes, dates, locations or ways to vote, or about voter eligibility.
• Synthetic depiction of candidates, officials or political figures saying or doing what they did not say or do, even if presented as satire, where there is a reasonable risk of it being taken as real.
• Political advertising that violates applicable electoral law, including, in Brazil, the rules of the TSE (Brazilian Superior Electoral Court) on the use of AI in electoral advertising (mandatory labeling and prohibition of deepfakes).

3.6. Infringement of intellectual property and personality rights

• Reproduction of copyrighted works, registered trademarks, characters or designs of third parties without authorization.
• Prompts that deliberately seek to imitate the style of an identifiable living artist in order to pass off the result as their work, or for commercially confusing purposes.
• Use of logos, packaging or visual identity of other brands in a way that suggests a non-existent affiliation.

3.7. Spam and Platform abuse

• Mass generation of content for spam, manipulation of search engines or platforms, or creation of fake account networks.
• Automation that circumvents limits, moderation or credit billing.
• Use of the API or MCP to systematically extract outputs for the purpose of training or evaluating competing models.

3.8. Other unlawful or high-risk uses

• Any content that is illegal in Brazil or in the country where you are located or to which you direct the content.
• Content that facilitates illegal activities (weapons, drugs, human trafficking, exploitation).
• High-risk automated decisions (credit, employment, health, justice) based solely on Trakto outputs; the Platform was not designed for this.
• Medical, legal or financial advice presented as professional without the supervision of a licensed professional.

4. Synthetic content and disclosure

4.1. When to disclose. You must indicate that content was generated or substantially altered by AI whenever: (a) applicable law requires it (for example, the European AI Act for deepfakes and content that interacts with people; Brazilian electoral rules; US state laws on synthetic content); (b) the destination platform requires it (social networks, ad marketplaces); (c) the content depicts real people or real events photorealistically and could be taken as an authentic record.

4.2. How to disclose. Use visible labels ("generated with AI", "synthetic image"), descriptions or metadata, as appropriate to the medium. Trakto may include metadata and invisible watermarks in outputs; do not remove them when the law requires that they be kept. provenance metadata inherited from providers (SynthID on Gemini/Imagen/Veo outputs), no proprietary C2PA

4.3. Artistic and obvious contexts. The labeling obligation generally does not apply to manifestly artistic, stylized or fictional content that no reasonable person would take as a real record, unless a specific legal requirement applies.

5. Image, voice and intellectual property rights

5.1. Real people. When you upload photos, videos or voices of real people, you represent that you have authorization from those people (or their legal guardians) for the intended use, including AI editing. Keep that authorization: we may request it in the event of a report.

5.2. Third-party works. When you use images, text, music or other third-party materials as reference or input, you represent that you have a license or other legal basis to do so.

5.3. Trakto templates and assets. They may be used as a basis for your designs, including with AI, in accordance with the Terms of Use.

5.4. Outputs. To the extent permitted by law, the outputs are yours. Keep in mind that copyright protection for AI-generated content varies between countries and that similar outputs may be generated for other people. Before sensitive uses (trademark registration, large-scale campaigns), do your due diligence.

6. How we moderate

6.1. Automated moderation. Prompts, inputs and outputs go through automated classifiers (Trakto's and the providers') that may block the generation, blur results or flag content for review. Automated filters err in both directions: they may block legitimate content and let prohibited content through. If you believe a generation was wrongly blocked, contact support.

6.2. Human review. Flagged content may be reviewed by trained members of the Trakto team, under confidentiality, with restricted and logged access, exclusively for moderation and safety purposes. We do not review content proactively and indiscriminately.

6.3. Reports. Anyone may report content published on Trakto Links or outputs that violate this Policy through abuse@trakto.io. We assess reports with priority proportional to their severity. For copyright infringement, we accept notices identifying the work, the infringing content, contact details and a good-faith statement; the notified user may submit a counter-notice.

6.4. Cooperation with authorities. We report child sexual exploitation content and other serious crimes to the competent authorities, as required by law.

7. Consequences of violations

7.1. Depending on severity, recurrence and context, Trakto may: (a) block the specific generation; (b) remove or unpublish content; (c) issue a warning; (d) temporarily restrict AI features; (e) suspend the account; (f) terminate the account without refund of credits; (g) refuse new purchases; (h) notify authorities.

7.2. Serious violations (sexual content involving minors, non-consensual intimate deepfakes, terrorism, organized fraud) result in immediate termination.

7.3. Appeal. You may appeal a moderation measure through support. We will review it with a person other than the one who applied the measure, within 10 business days, and inform you of the outcome with reasons.

7.4. Credits consumed in generations blocked for violating this Policy are not refunded. generations blocked by the provider's safety filter do not consume credits

8. Liability and indemnification

You are fully responsible for the content you create with Trakto, before third parties and authorities, and will indemnify Trakto for claims arising from your use in violation of this Policy, in accordance with the Terms of Use.

9. Updates

This Policy evolves together with technology and legislation. Material changes will be communicated at least 15 days in advance; additional restrictions necessary to comply with the law or protect people may take effect immediately.

10. Contact

Questions about this Policy: Avenida Álvaro Otacílio, 3731, Bloco Espanha – Edf. JTR, Sala 602, Jatiúca, Maceió/AL, 57036-850, Brazil Abuse reports: abuse@trakto.io

Version history