Skip to content

MinT Terms of Service

Mind Lab Toolkit

Service Provider: MINDAI PTE. LTD.

Version: v2.0

Published: July [ ] 2026

Effective: [ ] 2026

PLEASE READ CAREFULLY BEFORE USE. These MinT Terms of Service (the "Terms" or this "Agreement") form a legally binding agreement between you (an individual user or a corporate entity, "you" or "Customer") and 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 ("Mindai", "we", "us" or the "Service Provider"). By clicking "I Agree", creating an account, or accessing or using any feature of MinT, you confirm that you have read, understood, and agreed to be bound by these Terms in their entirety. If you do not agree to any provision of these Terms, do not use MinT.

SPECIAL NOTICE. Provisions in BOLD or ALL CAPS in these Terms address material rights and obligations, including limitations of liability, warranty disclaimers and dispute resolution. Please review them with particular care. If you are accepting these Terms on behalf of an entity, you represent that you have the authority to bind that entity.

Article 1 — Definitions

Unless otherwise expressly defined, the following terms have the meanings set forth below:

  • "MinT" or the "Service" means Mind Lab Toolkit, an AI platform provided by Mindai comprising two independent components: (a) the Training Platform — a developer- and enterprise-facing reinforcement learning (RL) model-training infrastructure and parameter platform, with output consisting of model parameters (including model weights, checkpoints, LoRA and other adapters, and evaluation reports); and (b) the Model Gateway — providing API access to Macaron model series and third-party open-source models (such as DeepSeek, GLM, etc.). The Service comprises the related websites, APIs, SDKs, dashboards, and documentation.

  • "You" or "Customer" means the individual developer, researcher, or corporate entity that registers, accesses, or uses MinT, including, where applicable, all employees and representatives authorized to use MinT under your account.

  • "Account" means the unique credentials issued to you to access and use MinT.

  • "API Key" means the credential issued by Mindai that authenticates programmatic calls to MinT.

  • "Customer Data" means all data you upload, submit, generate, or process through MinT, including Training Data, Training Outputs, training-task configurations, run logs, and inference inputs and outputs, but excluding Account Information.

  • "Training Data" means datasets you upload to the MinT Training Platform for the purpose of training or fine-tuning models, and forms part of the Customer Data.

  • "Training Outputs" means the artifacts you produce by running training jobs on the MinT Training Platform, including model weights, checkpoints, adapters (e.g., LoRA weights), and evaluation reports.

  • "Base Model" means a pre-trained AI model integrated with or supported by the MinT Training Platform, whether provided by Mindai or by third parties (such as open-source communities or other model vendors).

  • "Subscription Plan" means the service subscription form you select, which varies depending on which MinT component you use: (A) Training Platform Subscriptions — charged based on compute resource usage or committed subscription scale, including without limitation the free trial tier, pay-as-you-go developer tier, enterprise tier, dedicated/private deployment, and research-preview tier; and (B) Model Gateway Subscriptions — charged based on the model category invoked and call volume, including without limitation Macaron model series API calls (billed by input/output Token usage) and third-party open-source model API calls (billed by input/output Token usage), and other forms that Mindai may offer from time to time. The exact features, supported model list, usage limits, and prices of each Subscription Plan are set forth on the MinT pricing page or in an order document executed between you and Mindai.

  • "SLA" means the MinT Service Level Agreement, which sets out the availability commitment and service-credit remedies and is incorporated into this Agreement by reference.

  • "AUP" means the MinT Acceptable Use Policy, which sets out prohibited uses and compliance requirements and is incorporated into this Agreement by reference.

  • "Privacy Policy" means the MinT Privacy Policy, which describes how Mindai processes personal information and is incorporated into this Agreement by reference.

  • "Affiliate" means any entity that directly or indirectly controls, is controlled by, or is under common control with Mindai.

  • "Beta Features" means features, products, or services labeled "beta", "experimental", "preview" or similar.

Article 2 — The Service and Eligibility

2.1 Service Description

MinT is an AI platform for developers, research teams, and enterprise customers, comprising two independent components: (A) the Training Platform — a reinforcement-learning model training infrastructure and parameter platform, providing, among other things: (i) compute scheduling and orchestration of distributed training jobs; (ii) unified interfaces for LoRA, full-parameter fine-tuning, and other training methods; (iii) monitoring, logging, and evaluation of training jobs; (iv) inference and deployment helpers; and (v) related APIs, SDKs, and developer documentation. (B) the Model Gateway — providing API access to (i) Mindai's proprietary Macaron model series, and (ii) third-party open-source models deployed by Mindai (such as DeepSeek, GLM). The exact features, supported model list, and limitations of each component are described in the MinT website and product documentation.

SERVICE NATURE — MinT CONSISTS OF TWO INDEPENDENT SERVICES UNDER A UNIFIED PLATFORM BRAND: (a) TRAINING PLATFORM — a reinforcement-learning model training infrastructure and parameter platform for developers and enterprise customers. You may use the Training Platform to train, fine-tune, and evaluate your own models; upon completion, you determine how to deploy and use the models. The output of the Training Platform consists solely of model parameters and does not include any content disseminated to the public. With respect to the Training Platform, Mindai provides the model training infrastructure and compute resources, whose outputs are model parameters, and Mindai itself does not directly provide generative-AI content or algorithm-recommendation services to the public; accordingly, at the Training Platform level, Mindai does not constitute a generative AI service provider under the PRC Interim Measures for the Management of Generative AI Services or an algorithmic recommendation service provider under the PRC Provisions on the Administration of Algorithm-Based Recommendations of Internet Information Services. How a model trained on the Training Platform is subsequently deployed, to whom it is offered, and what content it generates is determined by you, and you are solely responsible for complying with all filing, security-assessment, registration, content-moderation, and other regulatory obligations. (b) MODEL GATEWAY — Mindai provides API access to two categories of models: (i) Mindai's proprietary Macaron model series; and (ii) third-party open-source models deployed by Mindai under their respective open-source licenses. You may invoke these models via API Key through the Model Gateway for your own AI business purposes. With respect to the Model Gateway, Mindai, as the API service provider, may constitute a service provider or technical service supporter under applicable jurisdictions (depending on the specific model, your application scenario, and regulatory interpretation). If you use Model Gateway Outputs to provide services to third parties, you independently bear all obligations of a service provider.

2.2 Eligibility

To register, access or use MinT, you must satisfy all of the following:

  • You are at least 18 years of age (or the age of legal majority in your jurisdiction, whichever is higher) and have full legal capacity, or you are a duly organized and validly existing legal entity acting through an authorized representative;

  • You are not subject to any applicable economic sanctions or export-control list (including, without limitation, the lists maintained by the United Nations, the U.S. Office of Foreign Assets Control (OFAC), the European Union, Singapore, or the People's Republic of China);

  • You are not located in a jurisdiction prohibited by applicable law or by Section 17 of these Terms; and

  • You will comply with all laws and regulations applicable to your registration, use of, and the jurisdictions in which the Service is provided to you.

2.3 Account Registration

You shall provide true, accurate, current and complete registration information and shall promptly update such information when it changes. You are responsible for all activities that occur under your Account. Mindai may require completion of identity or entity-verification procedures and may restrict access to certain features until verification is complete.

2.4 Account Security

You shall safeguard your Account credentials and API Keys. In particular, you agree to:

  • Notify Mindai immediately upon discovery of any unauthorized access to or security incident affecting your Account;

  • Bear responsibility for all activities under your Account whether or not authorized by you; and

  • Apply industry-standard security measures (including multi-factor authentication, periodic credential rotation, and least-privilege scoping of API Keys) to protect your access credentials.

2.5 No Use by Minors; Child Safety

MinT IS NOT DIRECTED TO OR INTENDED FOR USE BY MINORS. MinT is an enterprise- and developer-facing AI platform (comprising the Training Platform and Model Gateway) and is not a consumer product. 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). Mindai does not knowingly collect Personal Information from minors through the Service; if Mindai becomes aware that it has received Personal Information from a minor without valid guardian consent, Mindai will delete such information without undue delay.

You further represent and warrant that you will not use MinT to: (a) train, fine-tune, or evaluate any model whose principal intended use is to produce, distribute, or otherwise facilitate child sexual abuse material (CSAM) or any other content that sexualizes, endangers, or exploits minors; (b) train any model for the purpose of grooming, extorting, intimidating, or otherwise targeting minors; or (c) upload as Training Data any imagery or content depicting minors in a sexualized, abusive, or exploitative manner. Mindai reserves the right, without prior notice, to suspend or terminate any Account and to report suspected CSAM-related activity to competent law-enforcement authorities in accordance with applicable law. The specific prohibited uses relating to minors are further set out in Section 6 of the AUP, which is incorporated into this Agreement by reference.

Article 3 — Subscription, Fees, and Billing

3.1 Subscription Plans

Mindai offers Subscription Plans separately for the two components of MinT: (A) TRAINING PLATFORM SUBSCRIPTIONS — charged based on compute resource usage or committed subscription scale, including without limitation: (i) a free trial tier with usage and feature limits; (ii) a pay-as-you-go developer tier; (iii) a fixed-fee enterprise tier (with negotiated service levels); (iv) a private/dedicated deployment offering with isolated compute and dedicated support; and (v) a research-preview or research-collaboration tier (granted on a case-by-case basis). (B) MODEL GATEWAY SUBSCRIPTIONS — charged based on the model category invoked and call volume, including without limitation: (i) Macaron model series API calls (billed by input/output Token usage); (ii) third-party open-source model API calls (billed by input/output Token usage); and (iii) other forms that Mindai may offer from time to time. The exact features, supported model list, usage limits, and prices of each Subscription Plan are set forth on the MinT pricing page or in an order document executed between you and Mindai.

3.2 Usage, Quotas, and Rate Limits

Your use of MinT is subject to usage thresholds that vary by Subscription Plan: (A) for the Training Platform, thresholds may include GPU-hour quotas, concurrent-job limits, and storage caps; and (B) for the Model Gateway, thresholds may include API-rate limits, concurrent-request limits, and input/output Token throughput limits. Mindai may throttle, queue, suspend, or apply overage charges where you exceed your applicable thresholds. Mindai may make commercially reasonable adjustments to such thresholds to preserve service quality and to prevent abuse, with prior notice where reasonably practicable.

3.3 Payment Terms

Fees for paid Subscription Plans are billed in accordance with your selected billing cycle (monthly or annual). You authorize Mindai to charge your designated payment method on each renewal date. All fees are exclusive of any applicable VAT, GST, sales, use, or other indirect taxes, which you are responsible for. Where currency-conversion or tax differences arise, Mindai may make reasonable adjustments.

3.4 Refunds

Paid Subscription Plans are non-refundable as a matter of general policy. If you are dissatisfied with the Service and request a refund within fourteen (14) days following your initial purchase, Mindai may, at its sole discretion, provide a full or partial refund or a service credit. Usage-based charges, fees for private/dedicated deployments, and customized professional-services fees are non-refundable once incurred.

3.5 Price Changes

Mindai may revise pricing for new or renewing Subscription Plans by giving at least thirty (30) days' notice via email, the MinT dashboard, or a website notice. Pricing for an in-progress, prepaid term will not change. If you do not accept revised pricing, you may terminate the affected Subscription Plan within fourteen (14) days following the notice and Mindai will refund the unused, prepaid portion on a pro-rata basis.

3.6 Late Payment

If payment is not received within seven (7) days after the due date, Mindai may suspend your access to the Service. Restoring access may require payment of all amounts owed plus reasonable late charges or collection costs. Mindai may terminate Accounts with amounts overdue for more than sixty (60) days.

Article 4 — Acceptable Use

4.1 General Obligations

You agree to use MinT only for lawful purposes and in compliance with these Terms and the AUP. The AUP is incorporated into and forms part of this Agreement and has the same legal force as these Terms.

4.2 Summary of Prohibited Uses

You shall not use MinT to (the full list is set out in the AUP):

  • (a) Train or fine-tune any model intended to be used for unlawful purposes, including without limitation models intended to generate child sexual abuse material (CSAM), to weaponize biological/chemical/nuclear/radiological technologies, to operate offensive cyber weapons, or to enable mass surveillance or political repression;

  • (b) Train models intended to facilitate fraud, phishing, social engineering, or impersonation of identity;

  • (c) Train models intended to generate, at scale, illegal content, disinformation, or content for election manipulation;

  • (d) Scrape, crawl, or otherwise collect copyrighted data without authorization for use as Training Data;

  • (e) Circumvent or tamper with MinT's safety mechanisms, quota controls, billing controls, or content filters;

  • (f) Reverse engineer, decompile, disassemble, or otherwise attempt to extract MinT's source code, internal model weights, training algorithms, or scheduling logic;

  • (g) Conduct unauthorized security research, penetration testing, vulnerability exploitation, or service-disruption activities; or

  • (h) Engage in high-frequency automated scraping, abuse of API quotas, or attacks on MinT's infrastructure or other customers.

  • (i) Use Model Gateway Outputs or model responses obtained otherwise to train, develop, or distill AI models that compete with the Macaron model series or other Mindai products;

  • (j) Access or use MinT services to build, develop, or provide any product or service that competes with Mindai's products, or resell MinT services in any form;

  • (k) Automate or programmatically bulk-scrape or extract Model Gateway Outputs (other than reasonable API calls expressly permitted by this Agreement).

4.3 Enforcement

Mindai may, based on automated detection, customer reports, or regulatory action, take any of the following measures against violations of this Section or the AUP: (a) issue a warning for first violations; (b) impose throttling or restrict specific features for repeat or material violations; (c) suspend or terminate the Account for serious violations or refusal to remediate; and (d) report suspected criminal conduct to competent authorities.

Article 5 — Customer Data and Training Data

CORE COMMITMENT. ALL CUSTOMER DATA YOU UPLOAD TO MinT (INCLUDING TRAINING DATA, TRAINING OUTPUTS, AND INFERENCE INPUTS AND OUTPUTS) REMAINS WHOLLY OWNED AND CONTROLLED BY YOU. MINDAI WILL NOT USE CUSTOMER DATA TO TRAIN ITS OWN MODELS OR FOR ANY OTHER COMMERCIAL PURPOSE WITHOUT YOUR PRIOR EXPRESS AUTHORIZATION.

5.1 Ownership

You retain all right, title, and interest, including all intellectual property rights, in and to the Customer Data. Nothing in this Agreement transfers any ownership of, or any intellectual property rights in, the Customer Data to Mindai. You are solely responsible for the legality, accuracy, and non-infringement of the Customer Data.

5.2 Limited Processing License

You grant Mindai a limited, non-exclusive, non-transferable (except as permitted hereunder), revocable, worldwide license to: (a) store, transmit and process the Customer Data as necessary to provide MinT in accordance with your instructions; (b) carry out technical processing required for service availability, security, and compliance; and (c) carry out other processing strictly within the scope of your express instructions. Other than as set forth above, Mindai will not use Customer Data for any other purpose. In particular, Mindai will not use your Customer Data to train Mindai's own or any other customer's models without your prior, separate, and express consent.

5.3 Data Lawfulness

You represent and warrant that the Training Data and other Customer Data you upload to MinT:

  • Have a lawful provenance (owned by you, properly licensed to you, or in the public domain and lawfully usable for commercial purposes);

  • Do not infringe any third party's intellectual property, privacy, publicity, or other rights;

  • Do not contain any unlawful or otherwise prohibited content under applicable law;

  • Where containing personal information, have been collected with the data subject's valid consent or under another lawful basis; and

  • Where involving cross-border data transfers, comply with all applicable data-export requirements (including, without limitation, requirements under Singapore PDPA, the EU GDPR, and the PRC PIPL).

5.4 Retention and Deletion

You may delete your Customer Data at any time through the MinT dashboard. Following termination of your Account, Mindai will delete the Customer Data within thirty (30) days, except where: (a) required to be retained by applicable law; (b) needed to be retained for ongoing security-incident investigation or dispute resolution; or (c) the data has been anonymized for aggregate service-statistics or security-analytics purposes. Once the retention period ends or the retention basis ceases to apply, Mindai will delete or anonymize the data without undue delay.

5.5 Data Security

Mindai applies industry-standard technical and organizational measures to protect Customer Data, including encryption in transit (TLS 1.2 or above), encryption at rest, access controls, audit logging, and periodic security testing. Further detail is set out in the MinT Security & Compliance Whitepaper.

Article 6 — Training Outputs and Model Weights

CORE COMMITMENT. ALL TRAINING OUTPUTS YOU PRODUCE USING THE MinT TRAINING PLATFORM (INCLUDING MODEL WEIGHTS, LoRA ADAPTERS, AND CHECKPOINTS) ARE WHOLLY OWNED BY YOU. MINDAI ASSERTS NO OWNERSHIP IN, AND IMPOSES NO RESTRICTION ON THE COMMERCIAL USE OF, YOUR TRAINING OUTPUTS.

6.1 Ownership of Training Outputs

All Training Outputs that you produce using the MinT Training Platform, including without limitation model weight files, LoRA and other parameter-efficient adapters, checkpoints, training logs, and evaluation reports, are wholly owned by you, together with all intellectual property rights therein. You may freely use, reproduce, modify, distribute, and commercialize the Training Outputs without paying any additional fees to, or seeking any additional license from, Mindai.

6.2 Pass-Through of Base-Model Licenses

IMPORTANT — BASE-MODEL LICENSES STILL APPLY. If you use any third-party Base Model integrated with the MinT Training Platform (including without limitation open-source models such as Llama, Qwen or ChatGLM, or any commercial API models), your use of the resulting Training Outputs remains subject to that Base Model's original license. the MinT Training Platform is only the training tool; nothing in this Agreement enlarges or alters your rights or obligations under the Base Model's original license.

Before using any Base Model, you shall review and comply with its original license, including without limitation:

  • The Llama family — the Meta Llama Community License;

  • The Qwen family — the Tongyi Qianwen License;

  • Apache 2.0, MIT, BSD, or other open-source license obligations (including attribution and license-text reproduction);

  • The terms of service of any commercial API model vendor.

6.3 Mindai's Minimum-Necessary Access

To provide the Service (e.g., training-job monitoring, troubleshooting, billing reconciliation), authorized Mindai personnel may, under strict access controls, access metadata about your Training Outputs (such as file size, training duration, loss curves), but will not access the actual weight content of the Training Outputs unless: (a) you expressly request technical assistance that requires such access; (b) required by applicable law or competent authority; or (c) reasonably necessary to investigate a serious security incident.

Article 7 — Mindai's Intellectual Property

7.1 MinT Platform IP

The MinT platform, source code, training framework, scheduling algorithms, user interfaces, documentation, trademarks, brand identifiers, and all related intellectual property rights are owned by Mindai or its licensors. These Terms grant you only a limited, non-exclusive, non-transferable, revocable license to access and use MinT subject to these Terms.

7.2 Restrictions

Except as expressly authorized by these Terms, you shall not:

  • Reverse engineer, decompile, or disassemble the MinT platform, or otherwise attempt to extract its source code;

  • Copy, modify, distribute, sell, resell or commercially exploit the MinT platform itself;

  • Remove, modify, or obscure any copyright, trademark, or other proprietary notice on MinT; or

  • Use the trademarks "MinT", "Mind Lab", "Mindai", or any confusingly similar marks without prior written permission.

7.3 Feedback

Mindai may freely use any feedback, suggestions, or improvement ideas you provide regarding MinT, without any obligation or compensation to you. Mindai will not, however, use such feedback in any way that personally identifies you.

Article 8 — Privacy and Personal Information

Mindai's processing of your personal information is governed by the separate MinT Privacy Policy, which is incorporated into this Agreement by reference. Subject to applicable law, the principal categories of personal information collected by Mindai include: (a) Account registration information (name, email, phone, organization); (b) Service-usage information (login logs, API call records, training-job metadata); (c) payment information (handled by third-party payment processors); (d) device and network information (IP address, browser/device type); and (e) information you submit to customer support.

Mindai complies with applicable data-protection laws, including the Singapore Personal Data Protection Act (PDPA), the EU General Data Protection Regulation (GDPR), and the PRC Personal Information Protection Law (PIPL). You may exercise your rights of access, rectification, deletion, portability, and withdrawal of consent by contacting contact@mindlab.ltd.

Article 9 — Third-Party Services and Open-Source Components

9.1 Third-Party Services

MinT may be integrated with third-party services (including cloud-compute providers, monitoring tools, payment processors, and open-source model repositories). Mindai exercises reasonable care in selecting third-party providers but does not assume responsibility for the availability, quality, or security of such third-party services. Your use of third-party services through MinT may also be subject to the terms of those third parties. Mindai maintains a Subprocessors List which is published on the MinT website.

9.2 Open-Source Components

MinT contains or integrates open-source software components, each subject to its own open-source license. A complete list of open-source components and corresponding license texts is available in the MinT documentation.

Article 9-A — Model Gateway Services

9-A.1 Service Definition

Mindai provides API access to the Macaron model series and third-party open-source models via the Model Gateway. The specific functional scope, supported model list, rate limits, and pricing shall be as stated in the MinT official documentation.

9-A.2 IP of Macaron Models

The model weights, algorithmic architecture, and all related intellectual property of the Macaron model series are owned in their entirety by Mindai or its licensors. Your API calls to Macaron models constitute only a limited, non-exclusive, non-sublicensable, non-transferable usage license, solely for use in your lawful business in accordance with this Agreement; you obtain no ownership, distillation rights, reproduction rights, or derivative development rights through such calls.

9-A.3 Third-Party Open-Source Models

Third-party open-source models are deployed by Mindai based on their respective open-source license terms. Your access via the Model Gateway is subject to this Agreement; you are obligated to comply with any downstream user requirements under the applicable open-source licenses. Mindai makes no express or implied warranty as to the output quality, accuracy, compliance, or non-infringement of third-party open-source models. Where upstream rights, availability, or legal status of the open-source models change, Mindai reserves the right to adjust, suspend, or terminate the availability of such models on the Model Gateway accordingly.

9-A.4 Inputs and Outputs

You retain ownership of Inputs submitted via the Model Gateway. You represent and warrant that: (i) you hold all necessary rights over the Inputs; and (ii) the Inputs do not violate applicable law or infringe third-party rights. To the extent permitted by applicable law, Mindai assigns to you all rights, title, and interest it may have in the Outputs, for your use in your lawful business. You acknowledge that due to the probabilistic nature of generative AI, other users may obtain similar Outputs; Mindai makes no warranty of originality or uniqueness. You shall not represent to any third party that any Output was human-created rather than AI-generated.

9-A.5 Model Gateway Data Handling

Inputs submitted through the Model Gateway and Outputs generated therefrom shall be processed by Mindai as a processor pursuant to the MinT Data Processing Addendum. Mindai commits not to use your Inputs or Outputs to train, fine-tune, evaluate, benchmark, or otherwise develop or improve any AI model of Mindai or any third party, except where you have provided prior, separate, express, and revocable written consent. Specific processing rules are as set out in the DPA and the MinT Privacy Policy.

Article 10 — Confidentiality

"Confidential Information" means non-public information disclosed by one party to the other that is marked confidential or that, by its nature, should reasonably be understood to be confidential. This includes business strategy, technical materials, customer data, pricing information, and unannounced product features. The receiving party shall: (a) protect the Confidential Information with at least the same degree of care it uses to protect its own confidential information of like importance; (b) limit disclosure to its employees, Affiliates, agents and contractors who have a reasonable need to know and who are bound by equivalent confidentiality obligations; and (c) use the Confidential Information solely for the purposes of performing this Agreement.

Confidential Information does not include information that is: (a) publicly known; (b) independently developed; (c) lawfully obtained from a third party without restriction; or (d) required to be disclosed by law or by competent authority (provided that the disclosing party shall disclose only the minimum information necessary to comply and, where legally permitted, shall give the other party prior reasonable notice so that the other party may seek a protective order or other appropriate remedy).

The confidentiality obligations survive termination of this Agreement for three (3) years; obligations with respect to information constituting a trade secret survive for so long as such information remains a trade secret.

Article 11 — Service Availability and SLA

Mindai will use commercially reasonable efforts to provide MinT on a continuous and stable basis. For paid subscribers, Mindai provides a monthly availability commitment. The committed availability levels, the calculation methodology, and the service-credit remedies for failures to meet such commitments are set out in the separate MinT Service Level Agreement (the "SLA"), which is incorporated into this Agreement by reference.

Free-trial users and research-preview users are not entitled to the SLA. Downtime resulting from any of the following is excluded from availability calculations: (a) scheduled maintenance (Mindai will give at least seventy-two (72) hours' notice); (b) force majeure; (c) suspensions caused by your breach of these Terms or the AUP; (d) failures of your network environment or third-party services; or (e) the portion of any private/dedicated deployment for which you have undertaken self-operation.

Article 12 — Beta Features

Mindai may make available features labeled "Beta", "experimental", "preview" or similar ("Beta Features"). Beta Features are provided AS IS, without any express or implied warranty. Mindai does not commit to: (a) generally release Beta Features; (b) maintain or support Beta Features; or (c) preserve any data generated by Beta Features. Mindai may modify, withdraw or discontinue Beta Features at any time without prior notice.

Article 13 — Disclaimers

TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, MinT AND ALL OUTPUTS (INCLUDING TRAINING OUTPUTS) ARE PROVIDED "AS IS" AND "AS AVAILABLE". MINDAI EXPRESSLY DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, AND NON-INFRINGEMENT.

Mindai does not warrant, in particular, that:

  • Any model trained using the MinT Training Platform will achieve any expected level of accuracy, performance, or convergence;

  • MinT will be uninterrupted, error-free, or free from defects or vulnerabilities;

  • Pre-trained weights provided by Base Model vendors are free of defects or security vulnerabilities; or

  • MinT is fit for any particular business or commercial use case of yours.

You acknowledge and agree that AI model training is inherently uncertain, stochastic, and experimental. You shall conduct adequate testing, validation, and safety evaluation of any Training Output before deploying it in your products or services.

Article 14 — Limitation of Liability

14.1 EXCLUSION OF DAMAGES. TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, AND WHETHER UNDER CONTRACT, TORT (INCLUDING NEGLIGENCE), STRICT LIABILITY OR ANY OTHER LEGAL THEORY, MINDAI SHALL NOT BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, PUNITIVE OR CONSEQUENTIAL DAMAGES ARISING OUT OF OR RELATED TO YOUR USE OF, OR INABILITY TO USE, MinT. SUCH DAMAGES INCLUDE, WITHOUT LIMITATION, LOST PROFITS, LOSS OF BUSINESS, LOSS OF CONTRACTS, LOSS OF REVENUE, LOSS OF GOODWILL, LOSS OF PRODUCTION, LOSS OF ANTICIPATED SAVINGS, LOSS OF DATA, AND COSTS OF PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES, REGARDLESS OF WHETHER MINDAI HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

14.2 AGGREGATE CAP. IN NO EVENT SHALL MINDAI'S AGGREGATE LIABILITY UNDER THIS AGREEMENT EXCEED THE TOTAL AMOUNT YOU ACTUALLY PAID TO MINDAI FOR THE SERVICE THAT GAVE RISE TO THE CLAIM DURING THE TWELVE (12) MONTHS PRECEDING THE EVENT GIVING RISE TO THE CLAIM.

14.3 CUMULATIVE LIMIT. ALL CLAIMS ARISING OUT OF OR RELATING TO THIS AGREEMENT, WHETHER BROUGHT BY YOU OR YOUR USERS AND REGARDLESS OF THE NUMBER, TYPE, OR CLAIMANT OF SUCH CLAIMS, SHALL BE AGGREGATED AND SUBJECT TO THE LIMIT SET FORTH IN SECTION 14.2.

14.4 Exceptions. The foregoing limitations do not apply to: (a) Mindai's willful misconduct or gross negligence; (b) Mindai's breach of its express commitments under Article 7 with respect to your IP; (c) Mindai's specific indemnification obligations under Article 15; or (d) any liability that may not be limited or excluded under applicable law.

Article 15 — Indemnification

15.1 Customer Indemnification

You agree to indemnify, defend, and hold harmless Mindai, its Affiliates, and their respective employees from and against any third-party claims, losses, liabilities, and reasonable legal costs arising out of or related to: (a) your breach of these Terms or the AUP; (b) the breach of the lawfulness representations in Section 5.3 with respect to your Customer Data; (c) any model that you trained using the MinT Training Platform (including its subsequent deployment and any content it generates) violating any law or infringing the rights of any third party; or (d) your breach of obligations under the original license of any Base Model.

15.2 Limited IP Indemnification by Mindai (Enterprise Customers)

For enterprise-tier and above paid customers, if the unmodified MinT Training Platform service itself (i.e., the Training Platform component set out in Section 2.1, excluding the Model Gateway, Macaron Outputs, third-party open-source models and their outputs) infringes the intellectual property rights of a third party, Mindai will: (a) defend such claim; (b) pay the amount of any final judgment or settlement; and (c) at its option, modify the Service, procure for you a continuing license, or terminate this Agreement and refund prepaid amounts on a pro-rata basis. This obligation does not apply where: (i) you have modified MinT; (ii) you combined MinT with third-party components not approved by Mindai; (iii) you failed to install upgrades or patches when notified by Mindai; (iv) any claims arising from Model Gateway Outputs or third-party open-source model invocations; (v) infringement stemming from your Input content itself; (vi) your use with knowledge or reason to know of infringement; (vii) your use of third-party trademarks in Outputs or practicing a patented invention through Outputs; or (viii) any combination of the foregoing, or any other claim not fundamentally arising from the inherent characteristics of the unmodified MinT Training Platform service itself.

Article 16 — Term and Termination

16.1 Term

This Agreement enters into effect upon your first acceptance and continues in force until terminated in accordance with this Article.

16.2 Termination by You

You may close your Account at any time through the MinT dashboard, which will terminate this Agreement.

16.3 Termination by Mindai

Mindai may suspend or terminate this Agreement or your Account access:

  • Where you have breached these Terms or the AUP and have failed to cure within thirty (30) days of notice from Mindai;

  • Where you engage in unlawful activity;

  • Where your Account presents a security risk or is being abused;

  • Where required by applicable law or regulation; or

  • With respect to free-trial users, on thirty (30) days' notice for any reason; or

  • Any other circumstance that Mindai reasonably believes poses a material or urgent risk to the security or availability of the Service, or to the legitimate rights or interests of Mindai, other users, or any third party.

16.4 Effects of Termination

Upon termination: (a) you shall immediately cease using MinT; (b) you may export your Customer Data and Training Outputs during a thirty (30)-day post-termination export window; (c) data not exported within that window will be handled in accordance with Section 5.4; and (d) Articles 1, 5, 6, 7, 10, 13, 14, 15 and 17–21 survive termination.

Article 17 — Export Controls, Sanctions, and AI Technology Controls

17.1 General Compliance

You shall comply with all applicable export-control and economic-sanctions laws, including without limitation the Singapore Strategic Goods (Control) Act, the U.S. Export Administration Regulations (EAR) and OFAC sanctions programs, the EU Dual-Use Regulation, and the PRC Export Control Law. MinT is not made available to: (a) users located in comprehensively sanctioned countries or regions (including without limitation Iran, North Korea, Syria, Cuba, and the Crimea region); (b) individuals or entities listed on any applicable export-control or sanctions list (including those maintained by Singapore, the U.S. OFAC, the EU, the UN, and the PRC Ministry of Commerce); or (c) users that intend to use MinT for controlled nuclear, chemical, biological or missile-technology end uses.

17.2 AI Technology Export Controls

AI-SPECIFIC EXPORT CONTROLS. You acknowledge that model weights, training algorithms, and related technology may, in certain jurisdictions, constitute controlled items subject to export-control regimes for artificial intelligence or dual-use technologies. Without limiting the generality of Section 17.1, you shall not use the MinT Training Platform to train, fine-tune, export, re-export, transfer, or otherwise make available any Training Outputs or Base Model derivatives in violation of: (a) U.S. controls on advanced computing and AI-related technologies (including applicable EAR entity-list, foreign direct product, and end-use restrictions); (b) PRC Export Control Law and any implementing measures governing dual-use items, technology, or services (including any restrictions adopted by the PRC Ministry of Commerce in respect of artificial-intelligence technology); or (c) any other AI- or dual-use-specific export-control regime of Singapore, the EU, or any other applicable jurisdiction.

17.3 Downstream Responsibility

You are solely responsible for obtaining any export licenses, registrations, or governmental authorizations required for your subsequent transfer, deployment, or commercialization of Training Outputs, including those that may be triggered by incorporating the Training Outputs into products or services offered to users in other jurisdictions. Mindai provides the training infrastructure only and does not thereby assume, and expressly disclaims, any obligation to secure such downstream export authorizations on your behalf.

Article 18 — Complaints, Reports, and Intellectual Property Infringement Notices

18.1 General Complaints and Reports

If you believe that another user is violating these Terms or the AUP, that your rights have been infringed, or that MinT is being operated in an unlawful manner, you may file a complaint with Mindai through the following channels:

  • Email: contact@mindlab.ltd

  • Response time: Mindai will provide a substantive response within fifteen (15) business days of receiving the complaint; for complex matters, the period may be reasonably extended but in no event will exceed sixty (60) days.

If you are not satisfied with Mindai's handling of your complaint, you may request review by Mindai's legal lead. You may also seek recourse through the competent regulatory authorities.

If you believe that content transmitted, stored, or generated through MinT infringes your copyright, trademark, or other intellectual property rights, please submit a written notice to Mindai through the following:

  • Email: contact@mindlab.ltd

  • Mailing address: MINDAI PTE. LTD., 152 Beach Road, #11-05, Gateway East, Singapore 189721

  • Attn: General Counsel / Copyright Agent

For Mindai to process your complaint effectively, your written notice must contain all of the following: (a) a physical or electronic signature of the person authorized to act on behalf of the rights holder; (b) a description of the copyrighted or otherwise protected work you claim has been infringed (or a representative list, where multiple works are involved); (c) a description of where the allegedly infringing material is located within the MinT service, sufficient for Mindai to locate it; (d) your name (or organization name), mailing address, telephone number, and email address; (e) a statement that you have a good-faith belief that the disputed use is not authorized by the rights holder, its agent, or the law; (f) a statement, under penalty of perjury or the equivalent legal consequence for false statements under applicable law, that the information in your notice is accurate and that you are the rights holder or authorized to act on the rights holder's behalf.

Upon receipt of a notice satisfying the foregoing requirements, Mindai may (but is not obligated to) remove, disable, or restrict access to the material alleged to be infringing. Mindai reserves the right to terminate the accounts of repeat infringers where appropriate. You are responsible for the accuracy of the information in your notice; bad-faith or false notices that cause losses to Mindai or any third party shall subject you to liability for damages.

Article 19 — Governing Law and Dispute Resolution

19.1 Governing Law

This Agreement is governed by, and shall be construed in accordance with, the laws of the Republic of Singapore, without reference to its conflict-of-laws principles.

19.2 Negotiation First

Any dispute arising out of or in connection with this Agreement shall first be addressed through good-faith negotiation between the parties, commencing within thirty (30) days after one party gives the other written notice of the dispute.

19.3 Arbitration

FAILING SETTLEMENT, THE DISPUTE SHALL BE FINALLY RESOLVED BY ARBITRATION ADMINISTERED BY THE SINGAPORE INTERNATIONAL ARBITRATION CENTRE (SIAC) IN ACCORDANCE WITH ITS RULES THEN IN FORCE. THE SEAT OF ARBITRATION SHALL BE SINGAPORE, THE LANGUAGE OF ARBITRATION SHALL BE ENGLISH, AND THE TRIBUNAL SHALL CONSIST OF ONE (1) ARBITRATOR (OR THREE (3) ARBITRATORS WHERE THE AMOUNT IN DISPUTE EXCEEDS USD 250,000). THE AWARD SHALL BE FINAL AND BINDING ON THE PARTIES.

19.4 Interim Relief

Notwithstanding the above, either party may seek injunctive or other interim relief in any court of competent jurisdiction, in particular in cases of intellectual-property infringement or breach of confidentiality.

19.5 Class-Action Waiver

To the extent permitted by applicable law, you agree to bring claims only in your individual capacity and not as plaintiff or class member in any purported class or representative action.

19.6 Mandatory Consumer Protections

If applicable mandatory consumer-protection law in your jurisdiction grants you rights or remedies in addition to those set out in this Article, those mandatory rights are not displaced by this Section.

Article 20 — Modifications

Mindai may modify this Agreement from time to time. Material modifications will be notified at least thirty (30) days in advance via email, the MinT dashboard, or a website notice; non-material modifications (such as typographical corrections or wording improvements) may take effect immediately. Continued use of MinT after the notice period constitutes your acceptance of the modified Terms. If you do not accept the modifications, you may terminate this Agreement during the notice period and Mindai will refund the unused, prepaid portion on a pro-rata basis.

Article 21 — General Provisions

21.1 Entire Agreement

These Terms, together with the MinT Privacy Policy, the AUP, the SLA, and any order documents or supplemental agreements separately executed between you and Mindai, constitute the entire agreement of the parties with respect to MinT and supersede any prior or contemporaneous oral or written agreements.

21.2 Severability

If any provision of this Agreement is held invalid, illegal, or unenforceable, the remaining provisions shall continue in full force and effect, and the invalid provision shall be replaced by a valid provision that most closely reflects the original intent.

21.3 No Waiver

Mindai's failure or delay in exercising any right under this Agreement does not constitute a waiver of that right.

21.4 Assignment

You shall not assign any of your rights or obligations under this Agreement without Mindai's prior written consent. Mindai may assign this Agreement to its Affiliate or successor in connection with a corporate reorganization, merger, acquisition or sale of assets, without your consent, but with prompt notice to you.

21.5 Notices

Notices from Mindai to you may be given by: (a) email to the address you provided at registration; or (b) posting on the MinT dashboard or the MinT website. Notices from you to Mindai shall be sent to contact@mindlab.ltd.

21.6 Language

This Agreement may be made available in both English and Chinese. In the event of any inconsistency, for individual users the version in the language selected by you at registration shall prevail; for enterprise customers the language specified in the order shall prevail; in the absence of such specification, the English version shall prevail.

21.7 Force Majeure

Mindai shall not be liable for any failure or delay in performance caused by force majeure (including without limitation natural disasters, war, pandemic, government action, regulatory orders, critical-infrastructure failures, cyber attacks, or the interruption, termination, license change, or unavailability of third-party open-source models, upstream APIs, or other third-party services on which Mindai depends). The affected party shall give prompt notice to the other and take reasonable mitigation steps. If a force majeure event continues for more than sixty (60) days and materially affects the purpose of this Agreement, either party may terminate this Agreement by written notice without liability for breach.

21.8 Independent Contractors

The parties are independent contractors. This Agreement does not create any agency, partnership, joint venture, or employment relationship.

21.9 No Third-Party Beneficiaries

Except as expressly set forth in this Agreement, no third party has any rights under this Agreement under the Contracts (Rights of Third Parties) Act 2001 of Singapore or otherwise.

Article 22 — Contact Information

If you have any questions, comments, complaints, or suggestions regarding this Agreement, please contact Mindai through the following channels:

  • Service Provider: MINDAI PTE. LTD.

  • General contact: contact@mindlab.ltd

  • Legal contact: contact@mindlab.ltd

  • Data Protection Officer: contact@mindlab.ltd

  • Registered address: 152 Beach Road, #11-05, Gateway East, Singapore 189721

Acceptance

☐ I have fully read, understood, and agreed to all the terms of this MinT Terms of Service. I have noted in particular Article 13 (Disclaimers), Article 14 (Limitation of Liability), and Article 19 (Dispute Resolution).

☐ I am at least 18 years old and have full legal capacity to enter into this Agreement, or I am duly authorized to bind the entity on whose behalf I am accepting this Agreement.

☐ I agree that the MinT Privacy Policy, the AUP, and the SLA together form an integral part of this Agreement.

— End of Agreement —