MinT Acceptable Use Policy

Mind Lab Toolkit

Service Provider: MINDAI PTE. LTD.

Version: v1.0

Published: 23 April 2026

Effective: [ ] 2026

PLEASE READ CAREFULLY. This Acceptable Use Policy ("AUP") sets out the activities that are prohibited on MinT. It supplements, and is incorporated into, the MinT Terms of Service. By using MinT, you (the customer and any individual acting on your behalf) agree to comply with this AUP. Violation of this AUP may result in suspension or termination of your account, forfeiture of fees paid, and referral to law-enforcement or regulatory authorities.

MinT IS A TRAINING INFRASTRUCTURE AND PARAMETER PLATFORM. It produces only model parameters, weights, and related training artifacts; it does NOT, by itself, produce text, image, audio, video, or other generative content. The prohibitions in this AUP therefore focus on (a) the purposes for which you train models on MinT, (b) the data you upload for training, and (c) the manner in which you consume our compute resources. You remain solely responsible for any downstream deployment, distribution, or use of the models you train on MinT.

Article 1 — Scope and Definitions

This AUP applies to (a) all use of MinT, the MinT API, the MinT dashboard, and any related Mindai service; (b) all data uploaded to or processed through MinT; and (c) all model artifacts produced through MinT (including model weights, checkpoints, LoRA adapters, evaluation reports, and other Training Outputs).

Capitalised terms not defined in this AUP have the meanings given in the MinT Terms of Service. In addition:

"Training Use" means the purpose, design, or intended deployment of a model that you train, fine-tune, or evaluate using MinT.
"Training Data" means any dataset, prompt, sample, or other material that you upload to, or instruct MinT to access for the purpose of, training a model.
"Compute Abuse" means any activity that consumes Mindai compute resources for purposes other than the bona fide training, fine-tuning, evaluation, or sampling of machine-learning models.

Article 2 — General Principles

You must use MinT lawfully, responsibly, and in good faith. In particular, you must not:

Use MinT in any manner that violates applicable law, third-party rights, or this AUP;
Use MinT to harm, deceive, defraud, harass, surveil, or discriminate against others;
Use MinT to undermine the security, integrity, or availability of any system, network, or service;
Use MinT to circumvent the terms, safety mechanisms, or content controls of any third-party service, model, or dataset; or
Permit any third party to do any of the foregoing through your account.

Article 3 — Customer Responsibility

You are solely responsible for: (a) the Training Data you upload; (b) the Training Use of any model you train, fine-tune, or evaluate on MinT; (c) the Training Outputs you generate; and (d) any downstream deployment, distribution, or use of those Training Outputs by you or by any third party to whom you make them available. Mindai does not pre-screen, review, or approve your Training Data or Training Use.

Where you grant access to MinT to your employees, contractors, agents, or end users, you remain responsible for their acts and omissions as if they were your own. You must implement reasonable internal controls, including access controls, training, and monitoring, to ensure compliance with this AUP within your organisation.

Article 4 — Compliance with Law

You must comply with all laws and regulations applicable to you and to your use of MinT, including (without limitation):

Data protection and privacy laws (e.g., PDPA, GDPR, UK GDPR, PIPL, CCPA);
Intellectual property and trade-secret laws;
Sanctions, export-control, and anti-money-laundering laws of the United States, the European Union, the United Kingdom, the Republic of Singapore, the People's Republic of China, and other jurisdictions to which you or your activities are subject;
Sector-specific regulations applicable to your downstream use of Training Outputs (e.g., medical-device, financial-services, automotive, critical-infrastructure regulations);
AI- and algorithm-specific regulations applicable to you or to your end users (including, where applicable, regulations governing generative AI, deep synthesis, recommendation algorithms, automated decision-making, and AI risk management); and
Content-related laws (e.g., laws prohibiting child sexual abuse material, terrorist content, incitement, defamation, and hate speech).

Article 5 — Prohibited Training Uses

You must not use MinT to train, fine-tune, or evaluate any model whose intended purpose, foreseeable use, or actual deployment includes any of the following:

5.1 Child Sexual Abuse and Exploitation

Models intended or reasonably foreseeable to generate, depict, or facilitate child sexual abuse material (CSAM), the sexual exploitation of minors, or the grooming of minors.

5.2 Mass Violence and Weapons

Models intended or reasonably foreseeable to facilitate the development, acquisition, or deployment of chemical, biological, radiological, or nuclear (CBRN) weapons, or other weapons capable of mass casualties;
Models intended to plan, incite, or materially assist acts of terrorism or mass violence.

5.3 Malware and Offensive Cyber Operations

Models intended or reasonably foreseeable to generate malware, ransomware, exploits targeting specific systems, phishing kits, or to materially assist unauthorised access to information systems.

5.4 Targeted Harassment, Defamation, and Non-Consensual Intimate Imagery

Models intended or reasonably foreseeable to generate non-consensual intimate imagery (including non-consensual sexual deepfakes) of real persons;
Models intended to harass, defame, dox, or impersonate specific individuals.

5.5 Manipulation and Exploitation of Vulnerable Persons

Models intended or reasonably foreseeable to materially distort the behaviour of natural persons through subliminal, manipulative, or deceptive techniques in a manner that causes or is reasonably likely to cause significant harm; or
Models intended or reasonably foreseeable to exploit the vulnerabilities of specific groups (including children, persons with disabilities, and persons in vulnerable economic or social circumstances) in a manner that causes or is reasonably likely to cause significant harm.

5.6 Inference and Categorisation by Sensitive Personal Characteristics

Models intended or reasonably foreseeable to infer or categorise natural persons by sensitive personal characteristics (including race or ethnic origin, political opinions, trade-union membership, religious or philosophical beliefs, sex life, or sexual orientation) from biometric or behavioural data, except where expressly permitted by applicable law (e.g., lawful labelling or filtering of biometric datasets, or narrow law-enforcement use authorised by competent authorities); and
Emotion-recognition systems intended for deployment in workplaces or educational institutions, except where strictly necessary for medical or safety reasons and supported by a lawful basis.

Models intended or reasonably foreseeable to enable mass or untargeted surveillance, social scoring of natural persons leading to detrimental or unfavourable treatment in unrelated contexts, predictive policing of individuals based solely on profiling or personality traits, or real-time or post remote biometric identification of natural persons in publicly accessible spaces in violation of applicable law.

5.8 Circumvention of Safety, Copyright, or Content-Moderation Mechanisms

Models intended to bypass, disable, or circumvent the safety mechanisms, content-moderation policies, copyright protections, or terms of service of any third-party AI model, platform, or service (including, by way of example, jailbreaks of base models or watermark removal).

5.9 Election Manipulation and Coordinated Inauthentic Behaviour

Models intended or reasonably foreseeable to facilitate election fraud, voter suppression, the dissemination of materially false content about elections or candidates, or coordinated inauthentic behaviour designed to mislead the public on a matter of public concern.

5.10 High-Risk Autonomous Decision-Making Without Safeguards

Models intended for fully or substantially autonomous decision-making in high-risk domains (including medical diagnosis or treatment, financial credit or eligibility, employment or worker management, education or student assessment, criminal justice, law-enforcement, migration or border control, or access to essential public or private services) without (i) appropriate regulatory authorisation, (ii) meaningful human oversight, (iii) demonstrated accuracy and reliability for the intended use, and (iv) compliance with applicable AI risk-management requirements.

5.11 Critical Infrastructure

Models intended for use as a safety component in, or for the operational management of, critical infrastructure (including the supply of water, gas, heat, or electricity; transport networks; banking core systems; or digital infrastructure) without (i) sectoral regulatory authorisation, (ii) demonstrated safety assurance commensurate with the safety-critical nature of the application, and (iii) appropriate human oversight and fail-safe mechanisms.

5.12 Other Unlawful or Severely Harmful Purposes

Any other Training Use that is unlawful under applicable law or that, in Mindai's reasonable judgment, poses a substantial risk of severe harm to individuals, communities, or critical systems.

Article 6 — Minors and Child Safety

MinT is an enterprise- and developer-facing training infrastructure and is not directed to or intended for use by minors. Mindai treats child safety as a non-negotiable line and applies a zero-tolerance policy to any use of the Service that exploits, endangers, or sexualizes minors. This Article 6 supplements (and does not limit) Article 5.1, the Training Data Requirements set out in Article 7, and Section 2.5 of the Terms of Service.

6.1 No Use by Minors

You must not register for, access, or use MinT if you are under eighteen (18) years of age (or the age of legal majority in your jurisdiction, whichever is higher);
You must not enable, instruct, or permit any minor under your control or supervision to access, operate, or otherwise use MinT;
Where MinT is used by a corporate Customer, the Customer shall implement reasonable identity-verification and access-control measures sufficient to ensure that only authorized adult personnel access the Service.

6.2 Prohibited Training Uses Involving Minors

Without limiting Article 5, you must not use MinT to train, fine-tune, evaluate, or deploy any model whose principal intended use, foreseeable use, or actual deployment includes any of the following:

Generating, modifying, distributing, or facilitating the distribution of child sexual abuse material (CSAM), including any depiction (real, computer-generated, or otherwise synthetically rendered) that sexualizes or appears to sexualize a minor;
Grooming, sextorting, intimidating, deceiving, or otherwise targeting minors, whether for sexual exploitation, financial exploitation, radicalization, or self-harm;
Producing realistic non-consensual intimate imagery (including deepfakes) of any identifiable minor;
Inferring or categorizing minors by sensitive personal characteristics (including biometric, emotional, or psychological inferences) for behavioral targeting, advertising, or scoring;
Profiling minors for commercial advertising or for any decision having legal or similarly significant effects on the minor.

6.3 Prohibited Training Data Involving Minors

You must not upload to MinT, or otherwise process through MinT, any of the following Training Data:

CSAM or any other content the possession or distribution of which is prohibited under applicable child-protection laws (including, by way of example, U.S. 18 U.S.C. §§ 2251-2260, the EU Directive 2011/93/EU, the PRC Law on the Protection of Minors, and the PRC Regulations on the Protection of Minors in Cyberspace);
Sexualized, abusive, or exploitative imagery of minors, regardless of source or provenance;
Personal information of minors collected without verifiable guardian consent or another lawful basis under the applicable child-data-protection regime (e.g., COPPA in the United States, the GDPR (Article 8) in the European Union, the PRC Personal Information Protection Law (Articles 28 and 31), and the PRC Regulations on the Protection of Children's Personal Information Online);
Datasets enriched with biometric or behavioral identifiers of minors except where such collection has a documented lawful basis (e.g., medical research with appropriate ethics-board approval and parental consent).

6.4 Detection, Reporting, and Cooperation with Authorities

Mindai may apply automated and manual detection mechanisms (including hash-matching against known CSAM hash sets such as the NCMEC and IWF lists, and classifier-based screening) to the Service. Where Mindai becomes aware, through such detection, customer reports, or third-party notifications, of suspected CSAM or other content described in Article 6.3, Mindai will:

Preserve the relevant content and metadata as required by applicable law;
Report the matter to the U.S. National Center for Missing and Exploited Children (NCMEC) where required under U.S. law (18 U.S.C. § 2258A), and to other competent national authorities (including the Singapore Police Force and, where appropriate, PRC public-security authorities) as required under applicable law;
Suspend or terminate the offending Account immediately and without prior notice; and
Cooperate with law-enforcement and child-protection authorities in any subsequent investigation.

6.5 Reporting Channel

If you become aware of any actual or suspected violation of this Article 6 — including the presence of CSAM on, or the use of MinT to target minors through, the Service — you shall report the matter without undue delay to: contact@mindlab.ltd (with the subject line marked "CHILD SAFETY — URGENT"). Mindai will treat such reports with strict priority and will not retaliate against any person who reports a suspected child-safety violation in good faith.

Article 7 — Training Data Requirements

You represent and warrant that your Training Data:

Has been obtained from lawful sources, in compliance with applicable law, contractual obligations, and the terms of service of any platform from which it was sourced;
Does not contain child sexual abuse material, terrorist content, or other content the possession or distribution of which is prohibited under applicable law;
Where it contains personal data, has been collected and is being processed on a lawful basis (e.g., consent, performance of contract, legitimate interests, legal obligation), and you have provided all required notices to data subjects;
Where it contains sensitive personal information (as defined under applicable law), has been collected with the heightened consents or other legal bases required by applicable law;
Does not infringe any third party's intellectual property rights, trade secrets, publicity rights, or other proprietary rights, except to the extent expressly licensed to you;
Has not been obtained through unauthorised access to information systems, scraping in violation of platform terms, breach of confidentiality, or other unlawful means; and
Where required by applicable law (including PIPL, GDPR cross-border rules, and national-security/data-export regimes), has been transferred to MinT in compliance with such cross-border transfer requirements.

Mindai does not pre-screen Training Data and is not responsible for the lawfulness or content of Training Data. However, Mindai may, in response to a credible report or investigation, access, quarantine, or remove Training Data that it reasonably believes violates this AUP.

Article 8 — Compute Resource Abuse

You must not use MinT compute resources for any of the following:

Cryptocurrency mining, blockchain validation, or any proof-of-work or proof-of-stake operation;
Distributed denial-of-service (DDoS) attacks, the operation of botnets, or any other coordinated traffic generation directed at third-party systems;
Sending unsolicited bulk email, SMS, or other spam, including operating spam infrastructure;
Hosting, distributing, or routing illegal content, malware, command-and-control servers, phishing pages, or pirated content;
Operating an open proxy, anonymising relay, or other intermediary network service that materially obscures the identity of the actor causing harm to a third party;
Multiplying accounts, using stolen identities, sharing credentials, automating signup flows, or otherwise evading rate limits, free-tier limits, fair-use thresholds, or billing controls;
Reserving or hoarding compute capacity beyond the bona fide needs of your training, fine-tuning, evaluation, or sampling activities;
Performing benchmarking, capacity testing, or load testing of MinT itself without prior written authorisation from Mindai.

Article 9 — Security and Platform Integrity

You must not, and must not attempt to:

Probe, scan, penetration-test, or otherwise test the vulnerability of any Mindai system, network, or service, except pursuant to a vulnerability-disclosure program or written authorisation from Mindai;
Reverse engineer, decompile, or disassemble any non-customer-facing component of MinT, except to the extent expressly permitted by applicable law;
Bypass, circumvent, or disable any access controls, authentication mechanisms, isolation boundaries, or usage limits of MinT;
Interfere with, degrade, or attempt to gain unauthorised access to the data, sessions, training jobs, or compute resources of other Mindai customers;
Introduce malicious code, worms, trojans, or other harmful payloads into MinT or into systems with which MinT interacts;
Use MinT to develop, train, or evaluate a competing training-infrastructure product (this restriction does not apply to your internal development of models for your own use).

You must report any suspected security vulnerability in MinT to contact@mindlab.ltd promptly upon discovery and refrain from further exploitation pending coordinated disclosure.

Article 10 — Sanctions, Export Control, and Restricted Jurisdictions

You represent and warrant that:

You, your affiliates, your beneficial owners, and your end users are not on any restricted-party list maintained by the United States (including OFAC SDN, BIS Entity List, and Denied Persons List), the European Union, the United Kingdom, the United Nations, the Republic of Singapore, or any other applicable jurisdiction;
You are not located in, organised under the laws of, or ordinarily resident in any jurisdiction subject to comprehensive sanctions (currently including, without limitation, Cuba, Iran, North Korea, Syria, and the Crimea, Donetsk, and Luhansk regions of Ukraine);
You will not use MinT, or transfer Training Outputs produced on MinT, to any restricted party or to any restricted destination in violation of applicable sanctions or export-control laws;
You will not use MinT to develop, train, or distribute models whose export is restricted under any applicable export-control regime without first obtaining the required licences.

Mindai may screen accounts, transactions, and Training Outputs against applicable restricted-party lists and embargoed-destination lists, and may suspend or terminate accounts that fail such screening.

Article 11 — Third-Party Privacy and Intellectual Property

You must not use MinT to:

Train models on personal data of third parties without a lawful basis under applicable data-protection law;
Aggregate, correlate, or re-identify de-identified or pseudonymised personal data without authorisation;
Reconstruct, mirror, or substantially replicate proprietary datasets, model weights, or other intellectual property of third parties without authorisation;
Train models whose primary purpose is to circumvent the licensing terms of third-party datasets, models, or APIs (including by distillation, model stealing, or systematic prompt extraction).

Article 12 — Subprocessor AUP Pass-Through

MinT is delivered using infrastructure provided by third-party Subprocessors (including cloud, GPU, storage, and network providers). Such Subprocessors maintain their own acceptable-use policies that may apply to your use of MinT. By using MinT, you agree to comply with the acceptable-use policies of Mindai's Subprocessors, as published from time to time on the Subprocessor Page identified in the MinT Privacy Policy. Where a Subprocessor's acceptable-use policy is more restrictive than this AUP with respect to a particular activity, the more restrictive standard prevails as to that activity.

Article 13 — Reporting Violations

If you become aware of any actual or suspected violation of this AUP — whether by yourself, by another Mindai customer, or by a third party using Training Outputs produced on MinT — please report it promptly to contact@mindlab.ltd or, for general inquiries, to contact@mindlab.ltd. Reports should include, where available: (a) the nature of the suspected violation; (b) the identity of the party engaged in the violation; (c) the time and manner of the violation; and (d) any supporting evidence (e.g., URLs, screenshots, account identifiers). Mindai will treat reports confidentially to the extent reasonably practicable, consistent with the need to investigate and to comply with legal obligations.

Article 14 — Investigation, Suspension, and Termination

14.1 Investigation

Mindai may investigate suspected violations of this AUP. As part of an investigation, Mindai may (a) review account information, usage telemetry, and security logs; (b) where reasonable and lawful, access Training Data or Training Outputs (subject to the confidentiality and minimisation principles in the MinT Privacy Policy and the DPA); and (c) request information and cooperation from you.

14.2 Suspension

Mindai may suspend your access to MinT, in whole or in part, where Mindai reasonably determines that (a) a violation of this AUP is occurring or has occurred, (b) such suspension is necessary to protect the security, integrity, or availability of MinT or other customers, (c) such suspension is required by law or by an order of a competent authority, or (d) the situation otherwise constitutes a Grave Breach as defined in the MinT Terms of Service. Mindai will, where reasonably practicable and not prohibited by law, give prior notice of suspension and an opportunity to remediate; however, immediate suspension without prior notice is permitted in cases of severe or ongoing harm.

14.3 Termination

Mindai may terminate the MinT Terms of Service and your account in accordance with the termination provisions of the MinT Terms of Service, including for material or repeated violations of this AUP. Termination for AUP violation does not entitle you to a refund of fees already paid for services rendered.

14.4 Cooperation with Authorities

Mindai will cooperate with lawful regulatory, court, or law-enforcement requests relating to violations of this AUP, including by preserving and producing relevant records, in accordance with applicable law and the MinT Privacy Policy.

14.5 Preservation of Evidence

Where Mindai reasonably believes that records relating to a suspected violation may be required for an investigation or proceeding, Mindai may preserve such records (including Training Data and Training Outputs) beyond the standard retention periods set out in the MinT Privacy Policy until the investigation or proceeding is concluded.

Article 15 — Indemnification

You agree to indemnify, defend, and hold harmless Mindai, its affiliates, and their respective directors, officers, employees, and agents from and against any and all claims, losses, damages, liabilities, fines, penalties, and expenses (including reasonable legal fees) arising out of or in connection with any violation of this AUP by you, your personnel, or any party using Training Outputs produced on MinT, in accordance with the indemnification provisions of the MinT Terms of Service.

Article 16 — Updates to This AUP

Mindai may update this AUP from time to time to reflect changes in law, technology, threat landscape, or Mindai's business practices. Material updates will be notified in accordance with the MinT Terms of Service. Your continued use of MinT after the effective date of an update constitutes acceptance of the updated AUP.

Article 17 — Contact

Questions about this AUP, requests for clarification, or requests for written authorisation for activities that would otherwise be restricted should be directed to:

Service Provider: MINDAI PTE. LTD., a private limited company incorporated in the Republic of Singapore with its registered office at 152 Beach Road, #11-05, Gateway East, Singapore 189721
Email (abuse reports, security vulnerabilities, and general inquiries): contact@mindlab.ltd

Effective Date and Governing Versions

This AUP is effective from the date stated on the cover page. Mindai may publish localised versions of this AUP. In the event of any conflict between the English version and a localised translation, the English version shall prevail unless otherwise expressly required by applicable local law.

— End of Acceptable Use Policy —

On this page