THIS IS A LEGALLY BINDING AGREEMENT. IT CONTAINS BROAD DISCLAIMERS OF WARRANTY, A COMPREHENSIVE LIMITATION AND EXCLUSION OF LIABILITY, AN EXPRESS ASSUMPTION OF RISK BY YOU, AN INDEMNITY GIVEN BY YOU IN FAVOUR OF US, A RELEASE OF OUR DIRECTORS, OFFICERS AND PERSONNEL FROM PERSONAL LIABILITY, A WAIVER OF CLASS, GROUP AND REPRESENTATIVE PROCEEDINGS, AND A BINDING ARBITRATION CLAUSE. THESE PROVISIONS MATERIALLY LIMIT THE REMEDIES AVAILABLE TO YOU. IF YOU DO NOT ACCEPT THEM, YOU MUST NOT ACCESS OR USE THE SERVICE.
The Service is built on artificial intelligence. Artificial intelligence, and in particular the autonomous or semi-autonomous execution of tasks by software agents connected to third-party systems, is an emerging field. Its behaviour is probabilistic rather than deterministic. It is not fully predictable, not fully explainable, and not fully controllable — by us, by any model provider, or by anyone else at the present state of the art. Outcomes that no party intended, anticipated or could reasonably have foreseen are a known and accepted characteristic of this technology.
You use the Service entirely at your own risk. You are responsible for deciding whether the Service is appropriate for your purposes, for supervising what it does, for restricting the permissions you grant it, and for bearing the consequences of the actions it takes on your instruction or on your behalf.
A short summary follows. The summary is for convenience only and does not form part of the Agreement. Where the summary and the operative clauses differ, the operative clauses prevail.
1.1 In this Agreement, the following expressions have the following meanings.
"Agent" means any autonomous or semi-autonomous software process made available through the Service that receives instructions, plans a course of action, invokes Tools, calls Models, reads from or writes to Third-Party Systems, and produces Output — including without limitation the AI Worker suite marketed under the KOPI, Kopi O, Kopi C, Kopi Siew Dai, Kopi Gau, Kopi Po, Kopi Peng, Kopi Kosong and Kopi Human names, and any successor, variant, renamed or additional agent.
"Agreement" means these Terms of Service together with all documents incorporated by reference under clause 2.7.
"Agentic Action" means any act performed by an Agent that has effect outside the conversational context, including without limitation sending a message, creating, amending or deleting a record, issuing an instruction to a Third-Party System, initiating or approving a transaction, executing code, making a purchase, transmitting data, or omitting to do any of those things.
"Authorised User" means an individual whom you permit to access or use the Service under your account, whether an employee, officer, contractor, agent, affiliate, customer or other person.
"Company", "we", "us" or "our" means Kopi Ai Agent Pte Ltd, a private company limited by shares incorporated in the Republic of Singapore, together with its subsidiaries and affiliates where the context requires.
"Customer Data" means all data, content, documents, records, credentials, configuration and other material that you or any Authorised User submits to, stores in, or makes accessible to the Service, and all data that the Service retrieves from a Third-Party System on your instruction.
"Input" means any prompt, instruction, query, file, document, dataset, configuration, standing instruction, workflow definition or other material submitted to the Service by you or on your behalf, including by an Authorised User and including material submitted automatically by a system you have connected.
"Model" means any large language model, foundation model, embedding model, speech, vision, reasoning or other machine learning model accessed through the Service, whether operated by us, by a Model Provider, or by you.
"Model Provider" means any third party that develops, hosts, licenses or supplies a Model or inference capacity accessed through the Service.
"Output" means any text, code, data, image, audio, embedding, structured object, tool call, plan, recommendation, decision or other material generated or returned by the Service.
"Platform Channel" means any interface through which the Service is delivered, including our web application, dashboard, command line interface, SDKs, REST and streaming APIs, Model Context Protocol servers and connectors, e-mail, and the messaging platforms WhatsApp, Telegram and WeChat, and any successor or additional channel.
"Protected Persons" means the Company and each of its present and former directors (including non-executive and independent directors), officers, shareholders, investors, employees, secondees, contractors, consultants, advisers, agents, affiliates, licensors, suppliers, resellers, distributors and channel partners, and the respective successors and permitted assigns of each of them.
"Service" means the KOPI Agent platform in all its parts, including the Agents, the Token Desk, the Kopi TokenMax routing engine, the dashboard, the billing engine, the APIs, connectors and integrations, all documentation, and all associated software, infrastructure and support.
"Third-Party System" means any system, application, service, database, network, device or endpoint not operated by us that the Service connects to, reads from, writes to, or acts upon — including without limitation customer relationship management systems, enterprise resource planning systems, accounting and bookkeeping systems, banking and payment systems, e-mail and calendar systems, messaging platforms, file storage, ticketing systems, e-commerce platforms, government portals, and any system reachable by an application programming interface, webhook, browser automation or remote procedure call.
"Token" means a unit of computational measurement used by Models to quantify the volume of text or other data processed, and used by us as a basis for metering and billing. For the avoidance of all doubt, a Token is not a cryptocurrency, digital payment token, digital asset, security, unit in a collective investment scheme, e-money, stored value facility or any form of financial instrument, and confers no ownership, equity, profit-sharing, redemption or transfer right of any kind.
"Token Desk" means our wholesale procurement, aggregation, resale, metering and routing of Model inference capacity.
"Tool" means any function, connector, integration, plugin, script, browser automation or API binding that an Agent may invoke.
"you" or "Customer" means the person or entity accepting this Agreement, together with each Authorised User.
1.2 Headings are for convenience only and do not affect construction.
1.3 "Including", "includes" and "in particular" are to be read as "including without limitation" and do not limit the generality of the words that precede them.
1.4 A reference to a statute includes any subordinate legislation made under it and any amendment, consolidation or re-enactment of it from time to time.
1.5 The singular includes the plural and vice versa; a reference to a person includes an individual, body corporate, partnership, unincorporated association and government authority.
1.6 Where a provision excludes or limits liability, it is to be construed as excluding or limiting liability to the fullest extent that the applicable law allows, and no provision is to be read down further than that law requires.
1.7 The contra proferentem rule does not apply to this Agreement. This Agreement has been made available for review before acceptance, and you have had the opportunity to take independent legal advice.
2.1 This Agreement is formed when you first use the Service, by whatever means. You accept this Agreement, and it becomes binding on you, on the earliest of the following events:
2.2 Acceptance by conduct is sufficient. No signature, counter-signature, purchase order, written confirmation or formal execution is required. Your conduct in accessing or using the Service constitutes your unqualified acceptance of this Agreement in its entirety. If you do not agree to any part of this Agreement, your only remedy is to cease all use of the Service immediately.
2.3 Acceptance through informal channels is effective. You acknowledge and agree that interaction with the Service through consumer messaging platforms is informal in style but not informal in legal effect. A message sent to one of our Agents on WhatsApp, Telegram or WeChat has the same contractual consequence as a signed order form. You further acknowledge that these platforms are operated by third parties, that their user interfaces do not display these Terms, and that it is your responsibility to have read these Terms before initiating such an interaction. We direct you to these Terms at onboarding and at the address given in clause 35, and you agree that this constitutes reasonable notice.
2.4 Authority. If you accept this Agreement on behalf of a company, partnership, firm, government body or other entity, you represent and warrant that you have full authority to bind that entity, that you have obtained all internal approvals required, and that the entity is bound accordingly. If you lack that authority, you accept this Agreement in your personal capacity and are personally liable under it.
2.5 Capacity. You represent that you are at least eighteen years of age, or the age of legal majority in your jurisdiction if higher, and are legally capable of entering into binding contracts.
2.6 Authorised Users and downstream recipients. You are responsible for the acts and omissions of every Authorised User as if they were your own. You must ensure that every Authorised User is aware of, and complies with, this Agreement. Where you make the Service, or any Output, available to your own customers, clients, members or other third parties, you must impose terms on them that are at least as protective of the Protected Persons as this Agreement, and you remain fully liable to us for their conduct.
2.7 Documents incorporated by reference. The following are incorporated into and form part of this Agreement: our Acceptable Use Policy, Privacy Policy, Disclaimer, Open Source Attribution notice, Refund Policy, published product documentation, and any order form, statement of work, service schedule, data processing addendum or pilot agreement executed between us. In the event of conflict, the order of precedence is: (a) an executed order form or statement of work; (b) an executed data processing addendum, in respect of personal data only; (c) these Terms of Service; (d) the other incorporated documents. No terms contained in your purchase order, vendor portal, supplier onboarding pack, standard procurement conditions or any similar document have any effect, even if we sign, acknowledge or process such a document, unless we expressly agree to those terms in a written instrument that identifies this clause 2.7 and states that it overrides it.
2.8 No reliance. You acknowledge that you have not relied on any statement, representation, assurance, demonstration, benchmark, roadmap, marketing material, proposal, pitch deck, sales conversation or warranty other than those expressly set out in this Agreement. Nothing in this clause limits liability for fraudulent misrepresentation.
3.1 What we provide. The Service is a software platform that (a) makes Agents available across Platform Channels to perform work on your instruction; (b) procures, aggregates, meters, routes and resells access to Models through the Token Desk and the Kopi TokenMax routing engine; and (c) provides connectors, integrations and tooling by which Agents may interact with Third-Party Systems.
3.2 The Service is a tool, not a professional. The Service is not, and must not be treated as, a lawyer, accountant, auditor, tax adviser, financial adviser, investment adviser, insurance adviser, physician, engineer, licensed intermediary, fiduciary, employee, officer or agent of yours. It does not exercise professional judgment. It does not owe you a duty of care independent of this Agreement. It does not carry professional indemnity cover for your account.
3.3 The Service is probabilistic, not deterministic. Models generate Output by predicting likely continuations from patterns learned during training. Identical Input may produce different Output on different occasions. Output may be confidently expressed and entirely wrong. Output may cite sources that do not exist, state figures that are not real, and describe events that did not occur. This is an inherent property of the technology and is not a defect in the Service.
3.4 The Service operates at the frontier of a young field. Agentic artificial intelligence — the practice of allowing models to plan multi-step work, invoke tools, and act on external systems with limited or no human intervention at each step — is a new discipline. Established engineering standards, actuarial loss data, certification regimes, industry codes of practice and settled case law do not yet exist for it in any mature form. Failure modes are still being discovered. You acknowledge that you are deploying an emerging technology whose risks are not fully catalogued, and that neither we nor anyone else is in a position to enumerate them exhaustively in advance.
3.5 No completeness of warning. Any warning, risk description, guardrail, safety measure, filter, rate limit, approval gate or best-practice guidance we provide is offered in good faith on the basis of present knowledge. It is necessarily incomplete. The absence of a warning about a particular risk is not a representation that the risk does not exist, and does not transfer that risk to us.
3.6 We may change the Service. We may add, remove, modify, re-route, re-price, deprecate or discontinue any feature, Agent, Model, Tool, connector, Platform Channel or capability at any time, with or without notice, and without liability. See clauses 18 and 31.
4.1 You must provide accurate, current and complete registration information and keep it updated.
4.2 You are solely responsible for the security of your credentials, including API keys, bearer tokens, client secrets, OAuth grants, refresh tokens, webhook signing secrets, passwords, session cookies, multi-factor devices, recovery codes and any credential you supply to us for access to a Third-Party System.
4.3 All activity conducted through your credentials is deemed to be yours. This applies whether or not you authorised it, whether or not you were aware of it, and whether the activity was initiated by you, by an Authorised User, by an Agent, by an automated process, or by a third party who obtained your credentials by any means. You are liable for all Fees, all Token consumption and all consequences arising from such activity.
4.4 You must notify us at the address in clause 35 immediately on becoming aware of any actual or suspected unauthorised access, credential compromise, anomalous consumption or security incident. Until we have received that notice and had a reasonable period to act, all consequences remain yours.
4.5 You must implement and maintain reasonable security controls on your own systems and on the Third-Party Systems you connect, including least-privilege access, credential rotation, network restriction, logging and monitoring.
4.6 We may, but are under no obligation to, suspend an account, revoke a credential, throttle consumption or block a request where we reasonably suspect compromise, abuse, breach or risk to the Service or to other customers. We are not liable for any loss caused by such action, nor for any loss caused by our failure to take such action.
5.1 This clause is fundamental to the bargain between us. The fees payable for the Service are set on the express footing that you, and not we, bear responsibility for supervising the Service and validating what it produces and does. Absent that allocation, the Service could not be offered on the commercial terms on which it is offered.
5.2 You must, at your own cost and at all times:
5.3 Failure to perform any obligation in this clause 5 is a material breach of this Agreement, and is a complete defence to any claim you may bring against any Protected Person to the extent that the loss complained of would have been avoided or mitigated by performance of that obligation.
5.4 The analogy we ask you to hold in mind. The Service is a very fast, very capable and occasionally mistaken new hire. It works tirelessly, it does not sleep, and it will do exactly what it understood you to ask. It does not know what it does not know, and it will not tell you when it is wrong. You would not give a new hire on their first morning the keys to the safe, the company chop, unsupervised access to the banking portal and standing authority to e-mail your customers. You would start them on small tasks, check their work, and widen their authority as trust is earned. Do the same here. If you choose not to, the consequences are yours.
6.1 The central risk disclosure. This clause 6 describes the most significant category of risk associated with the Service. You are asked to read it with particular care. By using any agentic capability, you confirm that you have read, understood and accepted every part of it.
6.2 When you invoke an Agent, the Agent may, without a separate confirmation step at each stage: interpret your instruction; decompose it into sub-tasks; select and invoke Tools; call one or more Models; retrieve, create, amend, transmit or delete data in Third-Party Systems; issue instructions to those systems; iterate; retry; and continue until it determines the task complete or a limit is reached. These steps occur at machine speed. Many will complete before any human could intervene.
6.3 The specific sequence of steps an Agent takes is not fixed in advance, is not fully specified by us, may differ between runs with identical Input, and may not be reproducible after the event. The Agent may take a route to your objective that you did not contemplate and would not have chosen.
6.4 When an Agent acts on a Third-Party System using credentials or authorisations you have supplied, it acts as your instrument and on your authority, not ours. As between you and us, and as between you and any third party, every Agentic Action is deemed to be your own act, taken by you, with your knowledge, for your purposes and at your risk. We are a conduit and a tooling provider. We are not your agent, your fiduciary, your trustee, your employee, your delegate or your representative, and no such relationship arises by implication, custom, course of dealing or otherwise.
6.5 Standing instructions and unattended operation. If you configure recurring, scheduled, triggered, webhook-driven or otherwise unattended operation, you give a standing authorisation for the Agent to act without further reference to you, on each occasion, indefinitely, until you revoke it. You accept that unattended operation removes the principal safeguard available to you and that the risk of doing so is entirely yours.
6.6 Credential grants. Where you supply credentials, API keys, OAuth grants or delegated permissions for a Third-Party System, you represent and warrant that (a) you are entitled to grant that access; (b) the grant does not breach your agreement with the operator of that system or any other person; (c) you have obtained any consent required from data subjects, counterparties or licensors; and (d) you accept that any action performed with those credentials is attributable to you.
6.7 Many Agentic Actions cannot be undone. Once a message is sent, a record deleted, a payment initiated, a filing lodged, an order placed, a customer contacted, a file overwritten, a webhook fired or a piece of data transmitted, the effect may be permanent and beyond the power of any party to reverse. We have no ability to recall, rescind, unwind or remediate such an action. No rollback, undo, restore or compensation facility is offered or implied.
6.8 Without limiting the generality of the foregoing, you acknowledge that an Agent may, in the course of ordinary operation and without any defect in the Service, do any of the following in a manner you did not intend:
6.9 You accept full and sole responsibility for every consequence of every act and omission listed in clause 6.8 and for every analogous act or omission, whether or not listed.
6.10 Agents process content from documents, web pages, e-mails, messages, database fields, API responses and other sources. Content from any such source may contain instructions crafted to redirect the Agent. This is commonly called prompt injection. It may cause an Agent to disregard your instruction, exfiltrate data, take unauthorised action, or produce harmful Output. Related risks include poisoned or compromised tools and connectors, malicious Model Context Protocol servers, supply-chain compromise of an upstream dependency, and confused-deputy attacks that exploit the Agent's legitimate permissions.
6.11 No defence against these attack classes is complete. Mitigations exist and we apply reasonable measures, but the research community has not solved this problem and we do not represent that we have. You accept this risk. Where an Agent processes content that you do not control, you must assume that content may be adversarial and must restrict the Agent's permissions accordingly.
6.12 Agents may enter loops, retry excessively, fan out across large datasets, or recursively invoke themselves or one another. This may generate very large volumes of Token consumption, API calls, messages or writes in a short period. Fees and third-party charges so incurred are payable by you in full. Any cap, budget, quota or alert we make available is a convenience, not a guarantee, and may be exceeded before it takes effect. We do not warrant that any spending control will operate, operate in time, or operate accurately.
6.13 You expressly, knowingly and voluntarily assume all risk arising from the connection of Agents to Third-Party Systems and from the autonomous or semi-autonomous execution of Agentic Actions, including all risks described in this clause 6, all analogous risks, and all risks not yet identified or characterised by anyone. You agree that no Protected Person shall bear any liability whatsoever for any such risk that materialises, and you waive, to the fullest extent permitted by law, any claim in contract, tort (including negligence), equity, restitution, statute or otherwise in respect of it.
6.14 Controls we recommend and you agree to consider. Deploy first in a sandbox. Use read-only scopes until write access is demonstrably necessary. Require human approval for any action that moves money, contacts a customer, deletes data or is otherwise irreversible. Set hard spend caps and rate limits. Log everything. Keep a kill switch and test that it works. Segregate credentials by workflow. Review agent logs on a schedule, not only after an incident. These are recommendations, not warranties, and following them does not shift risk to us.
7.1 The Service is assembled, not self-contained. It depends on cloud and hosting providers, data centres, network carriers, submarine cable operators, Model Providers, inference and GPU capacity providers, messaging platform operators, payment processors, identity providers, open-source software maintainers, and other suppliers. We do not own, operate or control any of them.
7.2 We make no representation or warranty of any kind in respect of any third party, including as to availability, performance, latency, accuracy, security, data handling, continuity, pricing, policy, lawfulness or fitness for purpose. Your use of any Third-Party System is governed by your agreement with its operator, not by this Agreement.
7.3 Model substitution and routing. The Kopi TokenMax routing engine may direct a request to any Model in our supply chain, selected on grounds that may include cost, latency, capacity, context length, availability, jurisdiction and quality. Unless a specific Model is contractually fixed in an executed order form, we do not guarantee that any particular Model, version, provider, region or configuration will be used, will remain available, or will produce consistent results over time. Model Providers deprecate, retrain, re-align, re-price, rate-limit and withdraw models on their own schedules and without reference to us. Output characteristics may change materially as a result. This is not a defect and gives rise to no claim.
7.4 Messaging platform dependency. Delivery of Agents through WhatsApp, Telegram, WeChat and similar platforms is subject to those platforms' terms, policies, commercial arrangements, technical limits, rate limits, template approval regimes, message-window rules and enforcement decisions. Those platforms may restrict, suspend, throttle, ban or terminate access — to us, to you, or to a specific number, account or template — at any time, without notice, without reason and without recourse. Such an event does not constitute a breach by us, does not entitle you to any refund, credit or damages, and may occur even where neither you nor we have done anything wrong.
7.5 Regulatory dependency. Laws and regulations governing artificial intelligence, automated decision-making, cross-border data transfer, electronic communications, payments and platform liability are developing rapidly in Singapore and in every market we serve. A change in law, regulation, guidance or platform policy may require us to modify, restrict, geo-fence or withdraw any part of the Service, immediately and without compensation.
7.6 No third-party claims routed through us. If a third party's act or omission causes you loss, your recourse is against that third party. You will not bring, and hereby waive, any claim against any Protected Person founded on the act, omission, failure, outage, breach, insolvency, policy change or misconduct of any third party.
8.1 Metering. Token consumption is measured by our systems, by Model Providers' systems, or by a combination. Our records are conclusive evidence of consumption and of Fees payable, save in the case of manifest error demonstrated by you. Different Models tokenise text differently; the same text may yield materially different Token counts on different Models. Estimates, calculators and dashboards are indicative only.
8.2 Consumption you may not expect. Token consumption includes system prompts, tool definitions, retrieved context, conversation history re-sent on each turn, retries, reasoning traces, cache misses, failed calls and speculative work. Agentic and multi-turn workloads consume Tokens at rates far above single-shot use, because context is generally re-transmitted on every turn. You accept this and agree that consumption so incurred is chargeable.
8.3 No arbitrage or resale without consent. You may not resell, sublicense, redistribute, broker or provide onward access to Token capacity procured through the Token Desk except under an executed reseller or partner agreement with us.
8.4 Not a financial product. For the avoidance of all doubt, and repeating the definition in clause 1: Tokens and prepaid credits are units of account for computational services. They are not securities, digital payment tokens, cryptocurrency, e-money, stored value, deposits or investments. They confer no right to interest, return, redemption for cash, transfer to any other person, or participation in any profit. No representation is made that the purchase of credits is an investment or will appreciate in value.
8.5 Supply risk. Availability and price of Model capacity are outside our control and subject to global demand, hardware constraints, export controls and provider policy. We do not guarantee capacity, price stability, continuity of any supply relationship, or the ability to fulfil any particular volume.
9.1 Fees are as published on our website, as set out in your dashboard, or as agreed in an order form. Unless expressly stated otherwise, all Fees are exclusive of goods and services tax, value added tax, withholding tax and all other taxes, duties and levies, which are payable by you in addition.
9.2 Prepaid credits are non-refundable, except where a refund is required by applicable law or expressly provided in our Refund Policy. Credits may expire in accordance with the terms under which they were issued.
9.3 We may change pricing on thirty days' notice for subscription components. Consumption-based pricing and pass-through third-party costs may change with immediate effect where an upstream provider changes its pricing, and we will notify you as soon as reasonably practicable.
9.4 Invoices are payable within the period stated. Overdue amounts bear interest at 1.5 per cent per month or the maximum permitted by law, whichever is lower, and we may suspend the Service, revoke credentials and withhold data pending payment. You must pay all costs of recovery, including reasonable legal costs on a full indemnity basis.
9.5 No set-off. You must pay all sums due without set-off, counterclaim, deduction or withholding of any kind.
9.6 Disputed charges must be notified in writing within thirty days of the invoice date, with reasonable particulars. Charges not so disputed are deemed accepted and are final.
9.7 Chargebacks. Initiating a chargeback or payment reversal without first following clause 9.6 is a material breach and entitles us to suspend the Service immediately and to recover all resulting costs from you.
10.1 You must comply with our Acceptable Use Policy, which is incorporated into this Agreement. Without limiting it, you must not use the Service, and must not permit the Service to be used, to:
10.2 We may monitor use for compliance, security and billing purposes, may investigate suspected breaches, and may disclose information to law enforcement or regulators where required or where we reasonably consider it appropriate.
10.3 Breach of this clause 10 entitles us to suspend or terminate immediately, without refund, and you indemnify us in full for all consequences of such breach under clause 25.
11.1 The Service is not designed, tested, certified or warranted for use in any application where failure, error, delay, inaccuracy or unexpected behaviour could result in death, personal injury, severe environmental or property damage, loss of liberty, denial of a fundamental right, or catastrophic financial loss.
11.2 Without limitation, you must not use the Service as a sole or determinative input in, and you must not deploy it without qualified human review in:
11.3 If you use the Service in any high-risk field notwithstanding this clause, you do so entirely at your own risk, you represent that you have obtained all necessary regulatory approvals and professional oversight, and you indemnify each Protected Person in full and without cap under clause 25 in respect of all resulting claims, losses, penalties and costs.
11.4 Regulated activity. Nothing we provide constitutes the carrying on of any regulated activity for your account. If your use of the Service brings you within a licensing, registration, capital, conduct, disclosure or reporting regime, compliance is exclusively your responsibility. We do not hold ourselves out as licensed under the Financial Advisers Act 2001, the Securities and Futures Act 2001, the Payment Services Act 2019, the Insurance Act 1966, the Legal Profession Act 1966 or any equivalent legislation in any jurisdiction, and nothing in the Service should be construed as regulated advice or as a regulated payment, deposit-taking or dealing service.
12.1 Ownership. As between you and us, you retain all right, title and interest in Customer Data and Input. Subject to clause 12.4, and to the extent permitted by law and by the terms of the relevant Model Provider, we assign to you such rights as we have in the Output generated for you.
12.2 Licence to us. You grant us and our subprocessors a worldwide, non-exclusive, royalty-free licence to host, store, transmit, reproduce, display, adapt and process Customer Data, Input and Output solely to the extent necessary to provide, secure, meter, bill for, support and lawfully operate the Service.
12.3 We do not train on your content. We do not use your Input or Output to train, fine-tune or improve foundation models. However, you acknowledge that Model Providers operate under their own terms, that those terms vary, that they may change, and that we cannot audit or guarantee the internal practices of any third party. Where a specific data-handling commitment is material to you, it must be recorded in an executed order form or data processing addendum; otherwise no such commitment is given.
12.4 Non-uniqueness of Output. Output is generated statistically. Other customers may submit similar Input and receive similar or identical Output. We make no representation that Output is unique, original, novel, non-infringing, copyrightable, protectable, or free of third-party rights. Whether AI-generated material attracts copyright or other protection is unsettled in Singapore and in most jurisdictions. You must satisfy yourself as to the legal status of any Output you intend to rely on, register, publish or commercialise.
12.5 Accuracy of your Input. You warrant that you have the right to submit all Input, that its submission does not breach any law, contract, confidence or third-party right, and that you have given all notices and obtained all consents required.
12.6 Retention and deletion. Retention periods are set out in our Privacy Policy and product documentation. We may retain data as required for legal, tax, audit, security and dispute purposes. We do not warrant that deletion from the Service effects deletion from any Third-Party System, Model Provider log, backup, or cache outside our control.
13.1 Each party must keep confidential all non-public information of the other disclosed in connection with the Service, use it only for the purposes of this Agreement, and protect it with at least the degree of care it applies to its own confidential information and in any event no less than reasonable care.
13.2 The obligation does not apply to information that is or becomes public without breach, was lawfully known before disclosure, is lawfully received from a third party without restriction, or is independently developed. Disclosure compelled by law, regulation, court or regulator is permitted, with notice to the other party where lawful.
13.3 Do not submit information you cannot afford to have processed by third parties. Input is transmitted to Model Providers and other subprocessors. You must not submit legally privileged material, classified information, or information whose disclosure would be catastrophic, unless you have independently verified that the applicable processing chain is acceptable for that purpose.
14.1 Where we process personal data on your behalf, we act as a data intermediary and you act as the organisation with primary obligations under the Personal Data Protection Act 2012 of Singapore and any equivalent law applicable to you.
14.2 You are responsible for the lawfulness of your collection, use and disclosure of personal data, for consent and notification, for responding to data subject requests, and for conducting any impact assessment your regulator requires.
14.3 Cross-border transfer. Provision of the Service necessarily involves transfer of data across borders, including to jurisdictions whose data protection regimes differ from Singapore's. By using the Service you instruct and authorise such transfers. Where a specific data residency or transfer restriction is required, it must be agreed in an executed data processing addendum before use.
14.4 Nothing in this Agreement excludes or limits any obligation or liability that cannot lawfully be excluded or limited under applicable data protection law.
14.5 No security is absolute. We implement measures we consider appropriate, but no system is impenetrable. We do not warrant that the Service, or any data within it, will be free from unauthorised access, interception, corruption or loss, and, save to the extent liability cannot be excluded by law, we accept no liability for any security incident.
15.1 The Service, including the Kopi TokenMax routing engine, the Agent orchestration layer, the billing and metering engine, the dashboard, the connectors, all documentation and all associated know-how, is and remains our exclusive property or that of our licensors. No rights are granted except the limited, revocable, non-exclusive, non-transferable right to use the Service in accordance with this Agreement during its term.
15.2 Certain components are made available under open-source licences, identified in our Open Source Attribution notice. Those components are governed by their respective licences. Open-source components are provided by their authors without warranty, and we pass on no warranty in respect of them.
15.3 Trade marks. You may not use our names, marks, logos or branding without written consent, except to identify us factually as your supplier.
15.4 Feedback. Any suggestion, feature request, bug report, benchmark, comment or idea you provide is given freely, without obligation or expectation of compensation or confidentiality, and we may use and exploit it perpetually and without restriction.
16.1 You are on express notice that Output may be, and from time to time will be:
16.2 Output is not advice. It is not legal, medical, financial, investment, tax, accounting, engineering, safety, employment or professional advice of any kind, and no professional relationship arises from its provision. It does not represent our views, opinions, recommendations or endorsements.
16.3 No duty to warn. We are under no obligation to detect, flag, correct or notify you of any error in Output, and any confidence score, citation, source link, reasoning trace or caveat displayed is itself generated by a Model and is subject to the same limitations.
16.4 Reliance is at your risk alone. Any decision, action, omission, publication, transmission or commitment made in reliance on Output is your decision, and the consequences are yours.
17.1 The Service is provided on an "as available" basis. We do not guarantee, warrant or commit to any level of availability, uptime, latency, throughput, response time, capacity, accuracy, error rate or continuity.
17.2 No service level agreement applies unless a written service level agreement has been separately executed by an authorised signatory of the Company and expressly identifies this clause 17.2. Any availability figure, uptime percentage, latency target or performance statistic appearing on our website, in marketing material, in a proposal, in a pitch deck, in documentation, in a status page, or in any conversation is indicative and aspirational only, is not a contractual commitment, and is not capable of being relied upon.
17.3 We depend on parties we do not control. Availability of the Service is contingent on the availability, performance, capacity, policy and continued willingness to supply of, among others: cloud and hosting providers; data centre operators; power and cooling infrastructure; internet transit and peering providers; submarine cable operators; domain name and certificate authorities; content delivery networks; Model Providers and inference capacity providers; GPU and accelerator supply; messaging platform operators; payment processors; identity providers; and open-source maintainers. A failure, degradation, outage, rate limit, throttle, quota reduction, price change, policy change, suspension, withdrawal, insolvency or refusal to supply on the part of any of them will affect the Service, and we can neither prevent it nor compensate for it.
17.4 We strive but do not guarantee. We will use commercially reasonable efforts to maintain a high standard of service, to design for resilience, to monitor, and to restore service promptly after an interruption. That undertaking is one of effort and not of result. It creates no warranty, no guarantee, and no liability for any lapse, interruption, degradation, delay, data loss or failure of the Service, however caused and however serious.
17.5 No liability for downtime. Save to the extent liability cannot lawfully be excluded, no Protected Person shall be liable for any loss, damage, cost or expense arising from unavailability, interruption, degradation, delay, queuing, throttling, rate limiting, dropped connection, failed webhook, undelivered message, timeout or reduced capacity of the Service or any component of it. Any service credit, if offered, is your sole and exclusive remedy for such an event.
17.6 Scheduled and emergency maintenance. We may take the Service down for maintenance. We will endeavour to give reasonable notice for planned maintenance and to schedule it outside peak hours where practicable, but may act without notice where necessary for security, stability or legal compliance.
17.7 Backups. We may maintain backups for our own operational purposes. We do not provide a backup, archival or data recovery service to you, and you must not rely on us for the preservation or restoration of your data.
18.1 We may modify, improve, re-architect, re-route, re-price, restrict, degrade, deprecate, disable or discontinue any Agent, Model, Tool, connector, endpoint, Platform Channel, feature or the whole Service, at any time.
18.2 We will use reasonable efforts to give notice of material breaking changes to paid production features, but no minimum notice period is guaranteed and we may act immediately where required for security, legal, contractual, capacity or supply reasons.
18.3 Beta, preview, alpha, experimental, sandbox, trial and free-tier features are provided "as is", without any warranty, support, service level, data durability, backward compatibility or continuity commitment whatsoever, and may be withdrawn or changed without notice. Our aggregate liability in respect of any such feature is nil, and use of it is entirely at your own risk.
18.4 Deprecation of a feature does not entitle you to a refund, credit, damages or termination right, save where expressly provided in an executed order form.
19.1 We may suspend or restrict your access to all or part of the Service, immediately and without prior notice, where: (a) we reasonably suspect a breach of this Agreement or the Acceptable Use Policy; (b) we reasonably suspect fraud, credential compromise or unlawful activity; (c) your use threatens the security, stability, integrity, capacity or lawful operation of the Service or of any other customer; (d) payment is overdue; (e) an upstream provider, platform or regulator requires it; or (f) we are required to do so by law or court order.
19.2 We will lift a suspension when the cause is resolved to our reasonable satisfaction. We are not liable for any loss arising from a suspension made in accordance with this clause, and Fees continue to accrue during a suspension made under clause 19.1(a), (b), (c) or (d).
20.1 We maintain technical and organisational security measures that we consider appropriate to the nature of the Service. Details may be provided on request under confidentiality.
20.2 We do not warrant that the Service is secure, that it will remain secure, that it is free of vulnerabilities, malicious code or defects, or that unauthorised access will not occur. Threat actors are sophisticated, well resourced and persistent, and novel attack classes against AI systems are being discovered continuously.
20.3 You must not conduct security testing against the Service without our prior written consent. We operate a responsible disclosure process; contact details are in clause 35.
20.4 In the event of a security incident affecting your data, we will notify you without undue delay to the extent required by applicable law and will cooperate reasonably. Notification is not an admission of liability.
21.1 TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, THE SERVICE, THE AGENTS, THE MODELS, THE TOOLS, THE CONNECTORS, THE TOKEN DESK, THE DOCUMENTATION, ALL OUTPUT AND ALL RELATED MATERIALS ARE PROVIDED "AS IS", "AS AVAILABLE" AND "WITH ALL FAULTS", WITHOUT WARRANTY, CONDITION, REPRESENTATION, GUARANTEE OR TERM OF ANY KIND, WHETHER EXPRESS, IMPLIED, STATUTORY, COLLATERAL OR ARISING FROM COURSE OF DEALING, COURSE OF PERFORMANCE, USAGE OF TRADE OR OTHERWISE.
21.2 WITHOUT LIMITING CLAUSE 21.1, EACH PROTECTED PERSON EXPRESSLY DISCLAIMS ALL IMPLIED WARRANTIES AND CONDITIONS OF MERCHANTABILITY, SATISFACTORY QUALITY, FITNESS FOR A PARTICULAR PURPOSE, ACCURACY, RELIABILITY, COMPLETENESS, CURRENCY, TITLE, QUIET ENJOYMENT, NON-INFRINGEMENT, SYSTEM INTEGRATION, WORKMANLIKE EFFORT, AND ANY WARRANTY THAT THE SERVICE WILL MEET YOUR REQUIREMENTS, ACHIEVE ANY INTENDED RESULT, BE COMPATIBLE WITH ANY OTHER SOFTWARE OR SYSTEM, OPERATE WITHOUT INTERRUPTION, BE SECURE, BE ERROR-FREE, OR THAT ANY ERROR WILL BE CORRECTED.
21.3 NO PROTECTED PERSON WARRANTS THAT OUTPUT WILL BE ACCURATE, TRUE, COMPLETE, LAWFUL, NON-INFRINGING, UNBIASED, CONSISTENT, REPRODUCIBLE, ORIGINAL OR FIT FOR ANY PURPOSE; THAT ANY AGENTIC ACTION WILL BE CORRECT, TIMELY, AUTHORISED, REVERSIBLE OR CONSISTENT WITH YOUR INTENTION; THAT ANY THIRD-PARTY SYSTEM WILL BE AVAILABLE, COMPATIBLE OR CORRECTLY INTEGRATED; OR THAT ANY DATA WILL BE PRESERVED, RECOVERABLE OR FREE FROM CORRUPTION OR LOSS.
21.4 NO ADVICE OR INFORMATION, WHETHER ORAL OR WRITTEN, OBTAINED FROM ANY PROTECTED PERSON OR THROUGH THE SERVICE, CREATES ANY WARRANTY NOT EXPRESSLY STATED IN THIS AGREEMENT.
21.5 Some jurisdictions do not permit the exclusion of certain warranties. Where such a rule applies, the exclusions in this clause 21 apply to the maximum extent that law permits, and the remainder of the clause continues in full force.
22.1 You expressly, knowingly, voluntarily and irrevocably assume all risk arising from or connected with your use of the Service. This assumption of risk is a material inducement to us to make the Service available to you and forms part of the consideration for it.
22.2 Without limiting clause 22.1, you assume the risk of each of the following, whether or not caused or contributed to by the negligence of any Protected Person, and whether foreseeable or not:
22.3 You confirm that you have the resources, expertise, insurance and operational controls appropriate to bearing these risks, or have made an informed decision to proceed without them.
23.1 This clause 23 applies to all claims of every kind, whether in contract, tort (including negligence), breach of statutory duty, misrepresentation (other than fraudulent), restitution, equity, indemnity, strict liability or otherwise, and whether brought by you, by an Authorised User, or by any person claiming through or on behalf of you.
23.2 TO THE MAXIMUM EXTENT PERMITTED BY LAW, NO PROTECTED PERSON SHALL BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, EXEMPLARY, PUNITIVE OR AGGRAVATED LOSS OR DAMAGE, NOR FOR ANY OF THE FOLLOWING WHETHER DIRECT OR INDIRECT: LOSS OF PROFIT; LOSS OF REVENUE; LOSS OF ANTICIPATED SAVINGS; LOSS OF BUSINESS OR BUSINESS OPPORTUNITY; LOSS OF CONTRACT; LOSS OR CORRUPTION OF DATA OR SOFTWARE; COST OF DATA RECONSTRUCTION; LOSS OF GOODWILL OR REPUTATION; LOSS OF CUSTOMERS OR MARKET SHARE; MANAGEMENT OR STAFF TIME; BUSINESS INTERRUPTION; COST OF PROCUREMENT OF SUBSTITUTE SERVICES; REGULATORY FINES OR PENALTIES; OR LIABILITY TO ANY THIRD PARTY — IN EACH CASE HOWEVER ARISING, EVEN IF THE PROTECTED PERSON WAS ADVISED OF, KNEW OF, OR OUGHT REASONABLY TO HAVE FORESEEN THE POSSIBILITY OF SUCH LOSS.
23.3 TO THE MAXIMUM EXTENT PERMITTED BY LAW, THE TOTAL AGGREGATE LIABILITY OF ALL PROTECTED PERSONS TAKEN TOGETHER, FOR ALL CLAIMS ARISING OUT OF OR IN CONNECTION WITH THIS AGREEMENT OR THE SERVICE, SHALL NOT EXCEED THE LESSER OF (A) THE TOTAL FEES ACTUALLY PAID BY YOU TO THE COMPANY FOR THE SERVICE IN THE THREE MONTHS IMMEDIATELY PRECEDING THE FIRST EVENT GIVING RISE TO THE CLAIM, AND (B) ONE THOUSAND SINGAPORE DOLLARS.
23.4 The cap in clause 23.3 is an aggregate cap across all claims, all events, all causes of action and all Protected Persons. It is not a per-claim or per-event cap. A series of connected events constitutes a single event. Where the Service is provided free of charge, on trial, in beta or at no fee, the aggregate liability of all Protected Persons is nil.
23.5 Without limiting clauses 23.2 and 23.3, no Protected Person shall have any liability whatsoever, of any kind or in any amount, in respect of:
23.6 You acknowledge that the limitations and exclusions in this Agreement reflect a deliberate and negotiated allocation of risk between the parties; that they are reasonable having regard to the nature of the Service, the emergent state of the technology, the impossibility of insuring against unquantified risk, the price charged, and your ability to control your own exposure through the measures described in clauses 5 and 6.14; that the Fees would be substantially higher if we bore these risks; and that these limitations form an essential basis of the bargain and would apply even if any remedy is found to fail of its essential purpose.
23.7 You must take all reasonable steps to mitigate any loss. No Protected Person is liable for any loss to the extent caused or contributed to by your act, omission, breach, misconfiguration, instruction, failure to supervise, failure to test, failure to back up, failure to maintain fallback procedures, or by any third party.
23.8 Any claim must be brought within twelve months of the date on which the claimant first became aware, or ought reasonably to have become aware, of the facts giving rise to it. A claim not brought within that period is irrevocably barred and waived, notwithstanding any longer statutory limitation period, to the extent that such contractual shortening is permitted by law.
23.9 Nothing in this Agreement excludes or limits liability for: death or personal injury caused by negligence; fraud or fraudulent misrepresentation; any liability that cannot lawfully be excluded or limited under the Unfair Contract Terms Act 1977 of Singapore, the Consumer Protection (Fair Trading) Act 2003, the Personal Data Protection Act 2012, or any other applicable mandatory law. If any exclusion or limitation is held unenforceable, it shall be modified to the minimum extent necessary to make it enforceable, and all other provisions remain in full force.
24.1 The Company is a separate legal person. Your contract is with Kopi Ai Agent Pte Ltd alone. You agree that you will look solely to the assets of the Company for the satisfaction of any claim.
24.2 No claim against individuals. To the maximum extent permitted by law, no director (whether executive, non-executive, independent, alternate, nominee or shadow), officer, company secretary, shareholder, investor, employee, secondee, contractor, consultant, adviser, agent or affiliate of the Company shall have any personal liability of any kind to you or to any person claiming through you arising out of or in connection with this Agreement, the Service, any Output, any Agentic Action, any statement made in the course of marketing or negotiation, or any act or omission in the conduct of the Company's business.
24.3 Covenant not to sue. You covenant that you will not commence, join, fund, support or maintain any proceeding, claim, complaint or demand against any individual referred to in clause 24.2 in their personal capacity in respect of any matter falling within this Agreement. If you do so in breach of this covenant, you must indemnify that individual and the Company in full for all costs, expenses and losses incurred, on a full indemnity basis.
24.4 Non-executive and advisory roles. You specifically acknowledge that certain individuals associated with the Company act in a non-executive, advisory or strategic capacity only; that they are not involved in day-to-day operations, engineering, deployment, model selection or customer delivery; and that no representation, statement, introduction, endorsement or advisory contribution by any such individual creates any duty, warranty or liability owed by them to you.
24.5 Enforcement by third parties. Clauses 21, 22, 23, 24, 26 and 29 are intended to benefit each Protected Person. Each Protected Person who is not a party to this Agreement may enforce those clauses against you under the Contracts (Rights of Third Parties) Act 2001 of Singapore. Subject to that, no other person has any right to enforce any term of this Agreement, and the parties may vary or rescind this Agreement without the consent of any third party.
24.6 No piercing. You agree not to seek to pierce the corporate veil, to assert alter-ego, agency, joint venture, partnership, de facto director or accessory liability, or otherwise to attribute the obligations of the Company to any other person.
24.7 Nothing in this clause 24 purports to exclude any liability that an individual cannot lawfully exclude, including liability for fraud, or any duty owed to the Company itself under the Companies Act 1967 of Singapore.
25.1 You will indemnify, defend and hold harmless each Protected Person against all claims, demands, actions, proceedings, investigations, liabilities, judgments, settlements, awards, fines, penalties, losses, damages, costs and expenses (including legal costs on a full indemnity basis and the cost of internal management time) arising out of or in connection with:
25.2 This indemnity is uncapped and is not subject to clause 23. It survives termination.
25.3 Conduct of claims. We may, at our option, assume sole control of the defence and settlement of any indemnified claim using counsel of our choosing, at your cost. You must cooperate fully and must not admit liability, settle or compromise any claim without our prior written consent.
26.1 To the maximum extent permitted by law, you hereby release, waive, acquit and forever discharge each Protected Person from any and all claims, demands, causes of action, liabilities, losses and damages of every kind and nature, whether known or unknown, suspected or unsuspected, foreseen or unforeseen, present or future, arising out of or in any way connected with the Service, any Output, any Agentic Action, any interaction with any Third-Party System, or any matter in respect of which you have assumed risk under clause 22.
26.2 You expressly waive the benefit of any law, rule or principle that would otherwise limit the effect of a release to claims known at the time of release.
26.3 This release does not extend to any liability that cannot lawfully be released, including liability for fraud and for death or personal injury caused by negligence.
27.1 You are responsible for assessing whether your insurance arrangements are adequate for your intended use of the Service, including cover for cyber risk, technology errors and omissions, professional indemnity, business interruption, crime and regulatory defence costs.
27.2 We do not maintain insurance for your benefit, and no policy held by us confers any right on you. You must not represent to any insurer, regulator, auditor or counterparty that risk associated with your use of the Service is borne by us.
27.3 Where you are a regulated entity, you are responsible for satisfying your regulator's outsourcing, third-party risk, technology risk and business continuity requirements. We will provide reasonable information on request but give no assurance that the Service meets any particular regulatory standard.
28.1 This Agreement commences on your first acceptance under clause 2 and continues until terminated.
28.2 You may terminate at any time by ceasing use and closing your account. Termination does not entitle you to a refund of prepaid amounts, save as required by law or by our Refund Policy.
28.3 We may terminate (a) for convenience on thirty days' notice; (b) immediately for material breach; (c) immediately where required by law, regulator, court or upstream provider; (d) immediately where your use presents a security, legal, reputational or commercial risk we reasonably consider unacceptable; or (e) immediately on your insolvency, winding up, judicial management, receivership or comparable event.
28.4 On termination: all rights to use the Service cease immediately; outstanding Fees become immediately due; unused prepaid credits are forfeited save as required by law; and we may delete Customer Data after a reasonable period. You are responsible for exporting your data before termination. We give no undertaking to retain, restore or provide data after termination.
28.5 Disconnection risk. Termination or suspension will cause Agents to cease operating. Where you have made the Service part of an operational process, its cessation may disrupt that process. You must maintain the fallback procedures required by clause 5.2, and no Protected Person is liable for any consequence of cessation, however abrupt.
28.6 Survival. Clauses 1, 2.6, 2.8, 5.3, 6.4, 6.9, 6.13, 8, 9, 10.3, 12, 13, 14, 15, 16, 20.4, 21, 22, 23, 24, 25, 26, 27, 28.4, 28.5, 28.6, 29, 30, 33, 34 and 35, and any other provision which by its nature should survive, survive termination indefinitely.
29.1 No Protected Person is liable for any failure, delay, interruption or degradation in performance caused directly or indirectly by any event beyond our reasonable control, including: act of God; fire; flood; earthquake; storm; pandemic, epidemic or public health measure; war, hostilities, terrorism, civil unrest, riot or sabotage; act of government, regulator, court or authority; change in law, regulation, licence condition or official guidance; sanction, embargo or export control; strike, lockout or labour dispute; failure or shortage of power, cooling, water, telecommunications, internet transit, peering, submarine cable or satellite capacity; failure, outage, degradation, congestion, rate limiting, quota reduction, capacity constraint, price increase, policy change, deprecation, suspension, insolvency, acquisition or refusal to supply on the part of any cloud provider, data centre, hosting provider, Model Provider, inference or GPU capacity provider, messaging platform, payment processor, certificate authority, domain registrar, content delivery network, identity provider, open-source maintainer or other supplier; shortage of semiconductors, accelerators or computing capacity; cyber attack, denial-of-service attack, ransomware, malware, zero-day exploit, supply-chain compromise or unauthorised intrusion; defect or vulnerability in third-party software; and any other cause of a comparable nature.
29.2 The dependency on upstream providers described in clause 17.3 is expressly within the scope of this clause 29. A failure by any such provider is a force majeure event as between you and us, irrespective of whether it would be a force majeure event as between us and that provider.
29.3 Where a force majeure event continues for more than sixty consecutive days, either party may terminate on notice, without liability. Fees remain payable for services actually rendered.
30.1 You represent and warrant on a continuing basis that you, your Authorised Users, your beneficial owners and your end users are not the subject of any sanction, designation or restriction imposed by the United Nations, Singapore, the United States, the United Kingdom, the European Union or any other applicable authority, and are not located in or ordinarily resident in a comprehensively sanctioned territory.
30.2 You must not export, re-export, transfer or make available the Service, any Model access or any Output in contravention of any applicable export control, sanctions or dual-use regime.
30.3 Export controls applicable to advanced artificial intelligence models, model weights and computing hardware are subject to rapid and unpredictable change. A change may require us to restrict, geo-fence, suspend or withdraw access to any Model, capability, region or the Service as a whole, immediately and without compensation. Such action is not a breach of this Agreement.
30.4 You must comply with all applicable anti-bribery, anti-corruption, anti-money-laundering and counter-terrorism financing laws.
31.1 We may amend this Agreement at any time by publishing an updated version at the legal pages of our website and updating the effective date.
31.2 For material changes affecting paid production use, we will use reasonable efforts to give at least thirty days' notice by e-mail, in-product notice, or notice through a Platform Channel. Notice given through any Platform Channel, including a message on WhatsApp, Telegram or WeChat, is valid and sufficient notice.
31.3 Continued use of the Service after the effective date of an amendment constitutes acceptance of the amended Agreement. If you do not accept an amendment, your sole remedy is to cease use and terminate under clause 28.2.
31.4 Changes required for legal, regulatory, security or upstream-supplier reasons may take effect immediately.
31.5 It is your responsibility to review the current version periodically. We are not obliged to maintain or provide historical versions.
32.1 You consent to receive all communications, notices, agreements, disclosures and records electronically, and agree that electronic delivery satisfies any legal requirement for writing.
32.2 Notices to you may be given by e-mail to your registered address, by in-product notification, by posting to our website, or by message through any Platform Channel, and are deemed received on transmission or posting.
32.3 Notices to us must be sent to the address in clause 35 and are effective on acknowledged receipt.
32.4 Formal legal notices — including notices of claim, breach, dispute or termination for cause — must additionally be sent by registered post or courier to our registered office in Singapore.
33.1 Governing law. This Agreement and all non-contractual obligations arising out of or in connection with it are governed by and construed in accordance with the laws of the Republic of Singapore, without regard to conflict of laws principles. The United Nations Convention on Contracts for the International Sale of Goods does not apply.
33.2 Good faith negotiation. Before commencing any proceeding, the parties must attempt in good faith to resolve the dispute through discussion between senior representatives for a period of thirty days from written notice of the dispute.
33.3 Mediation. If negotiation fails, the parties must refer the dispute to mediation administered by the Singapore Mediation Centre before commencing arbitration, unless both parties agree in writing to dispense with mediation or urgent interim relief is required.
33.4 Arbitration. Any dispute not resolved under clauses 33.2 and 33.3 shall be referred to and finally resolved by arbitration administered by the Singapore International Arbitration Centre in accordance with its Arbitration Rules in force at the time. The seat shall be Singapore, the tribunal shall consist of one arbitrator, and the language shall be English. The award is final and binding. Judgment may be entered on the award in any court of competent jurisdiction.
33.5 Interim relief. Nothing prevents either party from seeking urgent injunctive or interim relief from the courts of Singapore, and the parties submit to the exclusive jurisdiction of those courts for that purpose and for the enforcement of any award.
33.6 Waiver of class, group and representative proceedings. To the maximum extent permitted by law, each party may bring claims against the other only in its individual capacity, and not as a claimant or class member in any purported class, collective, consolidated, group, representative or multi-party proceeding. No arbitrator or court may consolidate the claims of more than one person without the written consent of all parties. This waiver is a material term; if it is held unenforceable in a given case, clause 33.4 shall not apply to that case and it shall be determined by the courts of Singapore.
33.7 Confidentiality of proceedings. The existence, content and outcome of any dispute resolution process are confidential, save as required by law or for enforcement.
33.8 Costs. Unless the tribunal orders otherwise, the unsuccessful party bears the costs of the arbitration and the reasonable legal costs of the successful party.
34.1 Entire agreement. This Agreement, together with the documents incorporated under clause 2.7, constitutes the entire agreement between the parties and supersedes all prior discussions, proposals, representations, demonstrations, pitch materials and understandings.
34.2 Severability. If any provision is held invalid, illegal or unenforceable, it shall be severed or read down to the minimum extent necessary, and the remainder of the Agreement continues in full force.
34.3 No waiver. No failure or delay in exercising a right operates as a waiver. A waiver is effective only if in writing and signed, and applies only to the instance specified.
34.4 Assignment. You may not assign, novate, charge, subcontract or otherwise transfer this Agreement or any right under it without our prior written consent. We may assign, novate or transfer this Agreement freely, including in connection with a financing, reorganisation, merger, acquisition or sale of assets, and you consent in advance to any such transfer.
34.5 Subcontracting. We may use subcontractors, subprocessors, resellers, channel partners and affiliates in the performance of the Service.
34.6 No partnership or agency. Nothing creates a partnership, joint venture, agency, employment, franchise or fiduciary relationship.
34.7 Publicity. We may identify you as a customer and use your name and logo in customer lists and marketing materials, unless you notify us otherwise in writing.
34.8 Language. The English language version of this Agreement prevails over any translation. Translations are provided for convenience only and have no legal effect. Where a term is rendered into Chinese, the English term governs; in particular, the word "token" as used in this Agreement is a unit of computational measurement and is not to be read as a cryptocurrency or digital asset in any language.
34.9 Counterparts and electronic signature. Where any document is executed, it may be executed in counterparts and by electronic signature, each of which is an original.
34.10 Cumulative remedies. Rights and remedies under this Agreement are cumulative and not exclusive of any provided by law, save where expressly stated to be an exclusive remedy.
34.11 Further assurance. Each party will do all such acts and execute all such documents as may reasonably be required to give effect to this Agreement.
© 2026 Kopi Ai Agent Pte Ltd. All rights reserved.