Terms of Service

• "Services" — the LuluTokens gateway, Console, relay API, website, and the documentation at docs.lulutokens.ai.
• "Console" — the account dashboard at console.lulutokens.ai: keys, balance, usage logs, and settings.
• "Relay API" — the OpenAI-, OpenAI-Responses-, and Anthropic-compatible endpoints at api.lulutokens.ai.
• "API Key" — a credential (sk-…) issued from the Console that authenticates requests to the Relay API.
• "Model Provider" — a third party whose models we relay to, including those listed in the Console.
• "Input" — prompts, files, and any other content you transmit to the Services.
• "Output" — content returned by a Model Provider in response to your Input and relayed to you.
• "Customer Content" — Input, Output, and any other data you submit to or generate through the Services.
• "Credit" — prepaid balance held on your account and consumed as you use the Services.

LuluTokens sells prepaid API credit. You buy Credit in advance, in Singapore dollars, and each API request you make consumes Credit at the published rate for the model you called. There is no subscription, no recurring charge, and no minimum commitment. Nothing renews automatically.

LuluTokens is a relay. We provide:

• a single, standard-compatible API endpoint that reaches models operated by many different Model Providers, without integrating with each one separately;
• routing, channel scheduling, and failover between upstream channels;
• unified authentication, metering, quota enforcement, and billing across all Model Providers;
• a Console with usage logs, balance management, and account controls;
• documentation and support through the channels published on our website.

What we do not do. We do not train, own, host, or control the models themselves. We do not guarantee the accuracy, completeness, legality, or suitability of any Output. Beyond relaying your requests, we are not a party to the relationship between you and any Model Provider.

We may change the technical procedures, endpoints, model catalogue, routing behaviour, and supported features of the Services. We will use reasonable efforts to announce breaking changes to the Relay API in the documentation, but the model catalogue itself changes as channels are added and removed and is not guaranteed to be stable.

• You must register for an account to use the Relay API, and the information you give us must be accurate, complete, and kept current.
• You must be at least 18 years old, or the age of majority where you live, to hold an account. The Services are not directed to children.
• API keys are shown once. Keeping them confidential is your responsibility. Do not commit keys to source control and do not ship them in browser or mobile client code.
• All activity under your API keys is your responsibility, whether or not you authorised it, including all charges incurred. If you believe a key has been exposed, revoke it in the Console immediately and tell us at hello@lulutokens.ai.
• You may not share, resell, sublicense, or transfer your account or your API keys, or use them to operate a competing relay or resale service, unless we have agreed in writing.
• We may require identity verification, and may apply rate limits, spend limits, and model-level restrictions to an account for risk, compliance, capacity, or abuse-prevention reasons.

This section is important. Please read it.

• When you call a model, your Input is transmitted to the relevant Model Provider so that it can generate a response. That processing is governed by that provider's own terms, usage policies, and privacy practices, in addition to these Terms.
• You are responsible for ensuring that your use of a given model complies with that Model Provider's own usage policy. Where the two differ, the stricter rule applies.
• Model Providers may change their pricing, availability, rate limits, safety filters, and model versions at any time, and may deprecate or withdraw models without notice to us. We may consequently remove, reprice, or reroute models.
• Requests may be routed to any channel able to serve the requested model, including a channel in a different jurisdiction from your own, unless you have a written agreement with us that restricts routing.
• A model identifier names a model, not a supply chain. We do not guarantee that requests for a given identifier will always be served by the same upstream provider, infrastructure, region, or channel, and we may change any of those at any time without notice. If your workload requires a fixed provider or region, that needs a written agreement.
• We are not liable for a Model Provider's acts or omissions, including outages, refusals, content filtering, degraded quality, or their handling of your Input.

Your use of the Services is subject to the Usage Policy, which forms part of these Terms. In summary, and without limiting that policy, you may not use the Services to:

• carry out, promote, or facilitate illegal activity;
• generate or distribute content that is unlawful, that sexually exploits minors, that incites violence or hatred, or that infringes the rights of others;
• develop malware, attack any system, or circumvent the security or rate limits of the Services or of any Model Provider;
• generate deceptive content presented as human-authored where doing so is unlawful or materially misleading;
• resell, relay, or otherwise make the Services available to third parties as your own gateway, except under a written reseller or enterprise agreement; or
• use the Services in violation of any applicable sanctions, export-control, or trade-restriction law, or make them available to a person or place those laws prohibit.

If we find that your Customer Content or your use of the Services breaks the law, infringes a third party's rights, or breaches these Terms, we will notify you and may require the offending content or activity to be removed or stopped, and may disable access to it. Where the violation is severe, ongoing, or presents a legal or security risk, we may act first and notify you afterwards. If you do not remedy a notified violation within the period we specify, we may suspend or terminate the Services.

• You retain ownership of your Input. You grant us a limited, worldwide, non-exclusive licence to host, process, transmit, and display your Input solely as necessary to operate the Services, route your request to the relevant Model Provider, return the Output to you, meter usage, enforce this agreement, and comply with law.
• We do not use your Input or Output to train models, and we do not sell them. In normal operation we retain request metadata — timestamp, model, token counts, latency, status, cost, IP address — rather than the bodies of prompts and completions. See the Privacy Policy.
• Output. As between you and us, we claim no ownership of Output. Rights in Output are subject to the terms of the Model Provider that generated it and to applicable law, which in some jurisdictions may limit or deny copyright in machine-generated material.
• Output is not unique. Other customers may submit similar Inputs and receive similar or identical Outputs. Nothing here grants you exclusivity in Output.
• Output may be wrong. Outputs are generated statistically and may be inaccurate, out of date, offensive, or fabricated. You are responsible for evaluating Output before relying on it. Do not use Output as a substitute for professional advice in medical, legal, financial, or safety-critical settings without qualified human review.
• Our IP. The Services, the Console, our documentation, our trademarks, and the LuluTokens brand remain our property or that of our licensors. You may not copy, modify, reverse engineer, or create derivative works of the Services except as permitted by law.
• You must not remove, obscure, or alter any copyright, trademark, licence, or proprietary notice appearing in or on the Services or in any third-party software made available through them.

If you send us suggestions, bug reports, or other feedback, you grant us a perpetual, irrevocable, royalty-free licence to use it without restriction or obligation to you. We will not identify you as its source without your permission.

• You agree to pay for your use of the Services on the conditions set out in the Payment Terms and at the prices published in the Console, which are authoritative.
• Usage is metered per request and billed per token, separately for input and output tokens, at the rate applicable to the model and account group at the time of the request. Reasoning tokens, cached tokens, and other billable components are charged as described in the documentation and the Console.
• Charges are calculated from our metering records. Absent manifest error, those records are the authoritative measure of your usage. Billing disputes are handled as described in the Payment Terms.
• Taxes on the Fees, other than taxes on our net income, are payable by you. Our current tax status and how it affects the price you pay are set out in the Payment Terms.

We work to keep the Services available and will use commercially reasonable efforts to do so, but the Services are provided on an as-available basis and we do not offer a contractual uptime guarantee or service credits on prepaid accounts. If you need a committed availability target, an enterprise agreement can provide one — talk to us.

Scheduled maintenance is announced in advance where practicable. Emergency maintenance may happen with little or no notice.

Interruptions caused by a Model Provider — outages, rate limiting, content filtering, degraded quality, or latency at the provider — are the largest single source of failed requests on any relay, and are outside our control. So are events beyond our reasonable control, including network, utility, or infrastructure failure, and any change in law or regulatory action.

Failed requests are not billed. Where a request never reaches a Model Provider, no Credit is consumed. See the Payment Terms.

We may discontinue all or part of the Services on reasonable notice. If we permanently discontinue prepaid Services, our sole obligation is to do one of the following, at our election: (a) allow you to spend your remaining Credit during a wind-down period we announce; or (b) refund the unused balance where applicable law requires it. Nothing else is owed, and this is the only circumstance in which Credit is refunded.

We may suspend all or part of your access to the Services, with notice where practicable, if:

• your Credit is exhausted or a payment fails or is reversed;
• we reasonably believe your account is compromised, or is being used fraudulently, unlawfully, or in breach of the Usage Policy;
• your usage threatens the stability, security, or integrity of the Services, of a Model Provider, or of other customers — for example through abnormal request volumes, automated abuse, or attempts to bypass limits;
• a Model Provider, payment processor, or competent authority requires us to; or
• suspension is necessary to comply with law.

Access resumes as soon as the cause is resolved — for a prepaid account with no balance, as soon as a top-up clears.

• These Terms apply for as long as you hold an account.
• You may stop using the Services and close your account at any time. Read the Refund & Cancellation Policy first: closing an account forfeits any unused Credit, and Credit is not refundable in any case.
• We may terminate or suspend the Services if you breach these Terms in a way that cannot be remedied; breach them in a way that can be remedied and do not remedy it within the period we specify; become insolvent or cease to carry on business; or access the Services unlawfully, cause a security breach, or where we reasonably suspect fraudulent or unlawful activity.
• On termination, your right to use the Services ends immediately and your API keys stop working. Fees for Services already provided remain payable.
• After termination we retain and delete account and usage data as described in the Privacy Policy. Unused Credit is treated as set out in the Refund & Cancellation Policy.
• Any provision that by its nature should survive termination does so, including sections 4, 6, 8, 12, 13, 14, and 16.

The Services are provided "as is" and "as available", and your use of them is at your own risk.

To the maximum extent permitted by law, we disclaim all warranties, express or implied, including implied warranties of merchantability, satisfactory quality, fitness for a particular purpose, and non-infringement.

We do not warrant that the Services will be uninterrupted, timely, error-free, or completely secure, that any specific model will remain available, or that Output will be accurate, reliable, complete, non-infringing, or fit for your purpose.

This disclaimer applies to any damage, liability, or injury caused by failure of performance, error, omission, interruption, deletion, defect, delay in operation or transmission, computer virus, communication-line failure, or theft, destruction, unauthorised access to, or alteration or use of any record — whether framed as breach of contract, tort, negligence, or any other cause of action.

• To the maximum extent permitted by law, we will not be liable to you for indirect, consequential, exemplary, incidental, or punitive damages, including loss of profits, revenue, goodwill, data, or business, even if we have been advised of the possibility of such damages.
• We are not liable for any loss arising from your reliance on Output, from a Model Provider's acts or omissions, or from your failure to keep your own records of data you send to or receive from the Services.
• Our total aggregate liability arising out of or in connection with these Terms — whether in contract, tort (including negligence), by statute, or otherwise — is limited to the amounts received by us from you in the three (3) months immediately before the date the claim first arose.
• Nothing in these Terms excludes or limits liability that cannot lawfully be excluded or limited.

You will defend, indemnify, and hold us harmless against any claim, demand, loss, liability, or expense (including reasonable legal fees) brought by a third party and arising from: (a) your Customer Content; (b) your use of the Services or of any Output; (c) your breach of these Terms, the Usage Policy, or any Model Provider's terms; or (d) your violation of any law or of any third party's rights.

The gateway runs on infrastructure in Johor, Malaysia, and its database is hosted in Singapore. Payments are processed by Stripe, which operates globally. Model Providers process the Input you send them wherever they operate, which may be outside Singapore and Malaysia.

By using the Services you agree to this cross-border processing. What we collect, how long we keep it, and your rights over it are set out in the Privacy Policy.

We may access, preserve, and disclose logs, metadata, and limited request information where reasonably necessary to investigate abuse or a security incident, to detect or prevent fraud, to enforce these Terms, or to meet a legal or regulatory obligation. In normal operation we hold request metadata rather than the bodies of prompts and completions, and we do not routinely inspect content.

• Assignment. You may not assign these Terms without our written consent. We may assign them to an affiliate or in connection with a merger, acquisition, or sale of assets.
• No waiver. Failure to enforce a provision is not a waiver of it.
• Severability. If a provision is held unenforceable, the rest remains in force and that provision is modified to the minimum extent needed to make it enforceable.
• Entire agreement. These Terms, together with the Usage Policy, the Payment Terms, the Refund & Cancellation Policy, and the Privacy Policy, are the entire agreement between us on their subject matter. An executed enterprise agreement, where one exists, prevails to the extent of any conflict.
• Notices. We may give notice by email to the address on your account or by posting in the Console. You give notice to us at hello@lulutokens.ai.
• Third-party rights. A person who is not a party to these Terms has no right under the Contracts (Rights of Third Parties) Act 2001 of Singapore to enforce any of them.

These Terms are governed by the laws of Singapore, without regard to its conflict-of-laws rules.

If a dispute arises, the party claiming it must notify the other in writing with details. For thirty (30) days after that notice both parties will use best efforts to resolve it by negotiation. If it is not resolved in that period, the courts of Singapore have exclusive jurisdiction, and both parties submit to it.

Nothing in this section prevents either party from seeking urgent injunctive or other interim relief from any court of competent jurisdiction.

Matrix Minds Pte. Ltd. — operator of LuluTokens

• Email: hello@lulutokens.ai
• Address: 336 SMITH STREET, #05-301, NEW BRIDGE CENTRE, SINGAPORE (050336)

• Content you have the right to use. You may only submit content that you own, are licensed to use, or have otherwise lawfully obtained. This includes everything you put in a prompt: documents, code, images, and third-party or customer data.
• Follow the documentation. Use the Services in accordance with the technical documentation as it stands from time to time.
• Your own costs. All hardware, software, networking, and engineering resources needed to use the Services are your responsibility.
• Credential hygiene. You are responsible for keeping your API keys confidential and for all activity and charges incurred under them. Revoke a key in the Console the moment you suspect it has been exposed, and tell us.
• Accurate information. Registration details, billing details, and any information you give us during an investigation must be accurate, complete, and not misleading.
• Your own records. We retain request metadata, not the bodies of prompts and completions. If you need a record of what was sent or returned, you must log it yourself. We cannot reconstruct a conversation for you after the fact.
• Cooperation. You will provide information about Customer Content that we reasonably request in order to verify compliance with this policy, and will help us identify the source of any problem with the Services that we reasonably believe comes from your traffic.
• Downstream users. If you make the Services available to your own end users through your product, you are responsible for their conduct, for enforcing terms at least as protective as this policy, and for telling them that AI is being used and that its output may be inaccurate.
• Human review. Where Output influences decisions with legal, financial, medical, employment, safety, or similarly significant consequences, you must keep a qualified human in the loop. Do not present Output as professional advice.

You may not use the Services, or encourage, promote, facilitate, or instruct others to use them, for any illegal, harmful, or offensive purpose, nor to generate, transmit, store, display, or otherwise make available content that is illegal, harmful, or offensive. Prohibited activity includes:

Illegal activity

• Any unlawful activity, including operating or advertising unlicensed gambling services, trafficking in controlled substances or weapons, money laundering, or evading sanctions or export controls.
• Generating, soliciting, or distributing child sexual abuse material, or any material that sexually exploits or endangers a minor. We report this to the competent authorities and terminate the account without notice.

Harm to people

• Content that promotes, incites, or provides instructions for violence, terrorism, self-harm, or suicide.
• Content that harasses, bullies, defames, threatens, or intimidates an individual.
• Content that incites hatred or discrimination against a person or group on the basis of ethnicity, religion, race, nationality, gender, sexual orientation, disability, or any other protected characteristic.
• Generating sexual content involving real people without consent, non-consensual intimate imagery, or sexualised depictions of minors — under any framing.

Weapons and critical infrastructure

• Generating instructions or designs for chemical, biological, radiological, nuclear, or high-yield explosive weapons.
• Content intended to attack, disrupt, or compromise critical infrastructure — power, water, transport, telecommunications, healthcare, or financial systems.
• Use in life-support, nuclear, aviation, weapons-control, or other environments where failure could lead to death, personal injury, or severe physical or environmental damage.

Fraud and deception

• Offering or distributing fraudulent goods, services, schemes, or promotions; phishing; scam or social-engineering content; academic-fraud services; or fake reviews, engagement, or credentials.
• Impersonating a person, organisation, or government body without authorisation, including voice or likeness cloning without the subject's consent.
• Generating synthetic media presented as authentic, or coordinated political or election disinformation.
• Failing to disclose AI involvement where disclosure is required by law, or where a reasonable person would be materially misled by the omission.

Infringing and offensive content

• Content that infringes or misappropriates the intellectual property, privacy, publicity, or other proprietary rights of others.
• Content that is obscene, abusive, invasive of privacy, or otherwise objectionable, including content relating to bestiality or depicting non-consensual sex acts.

Harmful code

• Generating, refining, or distributing malware, ransomware, spyware, viruses, trojans, worms, or exploit code intended to damage, interfere with, surreptitiously intercept, or expropriate any system, program, or data. Authorised security research and defensive tooling are permitted; contact us if your workload is likely to trip automated detection.

Privacy and surveillance

• Building tools for mass surveillance, untargeted facial recognition, biometric categorisation by protected characteristic, social scoring, or covert tracking of individuals.
• Submitting special-category personal data (health, biometric, precise location, financial account, or government identifiers) belonging to third parties without a lawful basis and appropriate safeguards. If your workload requires it, tell us first — it may need a data processing agreement.

You may not use the Services to violate the security or integrity of any network, computer or communications system, software application, or computing device (each, a "System"). Prohibited activity includes:

• Unauthorised access — accessing or using any System without authorisation, including probing, scanning, or testing its vulnerability, or breaching any security or authentication measure.
• Interception — monitoring data or traffic on a System without permission.
• Falsification of origin — forging packet headers, email headers, or any part of a message that describes its origin or route.
• Attacking the gateway — attempting to access another customer's account, keys, balance, or logs; probing our infrastructure; or exploiting a defect in the Services rather than reporting it.
• Prompt injection against third parties — crafting Input designed to hijack, exfiltrate data from, or subvert another party's AI system or agent.

If you find a security vulnerability in LuluTokens, report it to hello@lulutokens.ai. Do not exploit it, do not access data that is not yours, and give us a reasonable period to fix it before disclosing it.

You may not make network connections to users, hosts, or networks that you do not have permission to communicate with. Prohibited activity includes monitoring or crawling a System in a way that impairs it, denial-of-service attacks, deliberate overloading, operating open proxies or traffic-laundering services, and using manual or automated means to evade any limit placed on a System.

The following apply specifically to LuluTokens:

• Do not evade rate or spend limits by rotating keys, creating multiple accounts, distributing traffic across accounts, or spoofing client identity.
• Do not abuse trial credit — one trial balance per person or organisation. Registering multiple accounts to collect trial credit is fraud and forfeits every balance involved.
• Do not resell or re-relay. You may not resell access, operate LuluTokens as the upstream of your own public API gateway, or provide the Services to third parties as though they were your own, without a written reseller or enterprise agreement with us.
• Do not extract the platform. Do not scrape the Console, reverse engineer the routing logic, systematically enumerate channels, or use Output to train, fine-tune, or distil a competing model, except where the relevant Model Provider expressly permits it.
• Do not circumvent model safety systems. Jailbreaking, adversarial prompting intended to defeat a provider's safety filters, and routing around model-level restrictions on your account are prohibited.
• Run a sane client. Use exponential backoff on 429 and 5xx responses; do not retry in a tight loop. Traffic patterns that threaten platform stability may be throttled or suspended without prior notice.

You may not distribute, publish, send, or facilitate the sending of unsolicited bulk email or other messages, promotions, advertising, or solicitations. You may not alter or obscure message headers or assume a sender's identity without that sender's explicit permission.

Using the Services to generate content for a spam, phishing, or bulk-solicitation campaign is prohibited even where the messages themselves are sent elsewhere.

Every request is served by a Model Provider under its own rules. You are responsible for complying with the policy of any provider whose models you call, as listed in the Console.

• Where this policy and a provider's policy differ, the stricter rule applies.
• A provider may block, filter, or refuse a request at its own discretion. That is not a fault in the Services and does not entitle you to a refund.
• If a provider requires us to suspend or restrict your access, we will comply.
• Provider terms may restrict what you may do with Output, including using it to train competing models. Those restrictions bind you.

• You must comply with all laws applicable to your use of the Services, including data protection, consumer protection, advertising, financial services, export control, and sanctions laws.
• Regional AI rules apply to you, not only to us. Where you serve users in a jurisdiction that regulates generative AI, you are responsible for meeting any labelling, filing, disclosure, record-keeping, and content-moderation duties that apply to your service.
• You may not export or re-export Output, or provide access to the Services, in violation of applicable export-control or sanctions regimes, and you may not use the Services if you are located in, or ordinarily resident in, an embargoed territory, or if you are a designated party under an applicable sanctions list.

To keep the platform running, we monitor operational metadata — request volumes, error rates, latency, models called, token counts, spend, and source IP. We do not routinely inspect the content of prompts or completions, and we do not retain them in normal operation. Where a violation is suspected or an investigation is legally required, we may inspect the traffic reasonably necessary to establish what happened.

We may:

• investigate suspected violations of this policy or other misuse of the Services;
• restrict, remove, disable, throttle, suspend, or terminate access to any account, key, model, or resource involved in a violation;
• report activity we suspect violates any law to law-enforcement bodies, regulators, or other appropriate third parties, including relevant customer information, and cooperate with them.

We generally give notice and an opportunity to remedy first. Where a violation is severe, ongoing, unlawful, or presents a security risk, we act immediately and notify afterwards. Suspension or termination under this section does not entitle you to a refund, and does not relieve you of Fees already incurred.

If you become aware of a violation of this policy — by another user, by content generated through the Services, or by your own systems — tell us at hello@lulutokens.ai and help us stop or remedy it.

Please include the request IDs, timestamps, model, and account or key involved where you have them. It makes an investigation dramatically faster.

We may update this Usage Policy as the platform, the model catalogue, and the law change. Material changes will be announced in the Console or by email before they take effect. Continuing to use the Services after a change takes effect means you accept it.

You are buying prepaid API credit — a balance denominated in Singapore dollars that is consumed as you send requests through the Relay API. It is not a subscription, not a licence, and not a right to any particular model remaining available.

Credit is sold in units of one Singapore dollar: S$1 buys one unit of credit. Your balance and every price in the Console are shown in Singapore dollars.

• All prices are in Singapore dollars (SGD). Stripe charges your payment method in SGD. If your card is denominated in another currency, your bank or card issuer sets the conversion rate and may add its own fees; we do not control or reimburse those.
• The price list in the Console is authoritative. Per-model prices are published per million tokens in the Console. Prices quoted anywhere else — including on this website and in the documentation — are illustrative.
• A single top-up is a minimum of S$5 and a maximum of S$500. Preset amounts within that range are offered in the Console, and you may top up as often as you like. The amounts available may change.

• Metering. Every request to the Relay API is metered and billed per token, separately for input and output tokens, at the rate for the model and the account group that applied at the moment of the request.
• What counts as billable tokens. Input tokens, output tokens, and — where the model produces them — reasoning tokens are all billable. Cached input tokens, where a model supports caching, are billed at the reduced rate shown in the Console.
• Deduction. The cost of each request is deducted from your balance when the request completes. Every response includes a usage object, and the Console's usage logs show the exact cost, model, token counts, and time of every individual request.
• Failed requests. Requests that fail before reaching a Model Provider are not billed. Requests that reach a provider and are billed to us are billed to you, including requests you cancel mid-stream.
• Our records govern. Charges are calculated from our metering records. Absent manifest error, those records are the authoritative measure of your usage — not token counts produced by third-party tokenizers or client-side estimates.

A completed top-up is required before the Relay API will serve any request. This includes models published at zero price: a free model costs nothing per token, but it is not a free tier, and it is served only to accounts that have paid.

• Any trial, promotional, or gift Credit we grant has no cash value, is never refundable, is not transferable, and may expire or be withdrawn at our discretion.
• One trial balance per person or organisation. Creating multiple accounts to collect additional trial credit breaches the Usage Policy and forfeits every balance involved.
• Higher rate limits and certain features may require a larger balance or a specific account group.

• Top-ups are made from the Console and are processed by Stripe. The payment methods available at checkout are those Stripe offers for your region and currency; they may include card payments and PayNow.
• Your payment details are collected and stored by Stripe under Stripe's own terms and privacy policy. We never receive or store your full card number.
• Credit is added to your balance once Stripe confirms the payment. Card payments are normally near-instant; methods that settle asynchronously may take longer to clear, and Credit appears only after the payment confirms.
• Stripe issues a payment receipt by email. Your top-up history is also available in the Console.
• The charge appears on your statement as LuluTokens. If you do not recognise a charge, email hello@lulutokens.ai before disputing it with your bank — see the Refund & Cancellation Policy.
• Credit can only be used to pay for LuluTokens usage. It cannot be transferred between accounts, exchanged for anything else, or converted back into money. See the Refund & Cancellation Policy.
• You are responsible for keeping enough Credit to cover your workload. We recommend enabling the low-balance alert in the Console.

Matrix Minds Pte. Ltd. is not currently registered for Singapore GST, so no GST is added at checkout: the amount shown in the Console is the amount Stripe charges. If we become GST-registered, tax will be shown at checkout before you confirm the payment.

Any other value-added, consumption, withholding, or similar tax levied on the Fees in your own jurisdiction is payable by you, other than taxes on our net income. Where you are required to withhold tax from a payment to us, you must gross up so that we receive the full amount.

If you need an invoice addressed to a registered entity, email hello@lulutokens.ai with the entity name, address, and any tax identification number.

• When your Credit is exhausted, API requests are rejected with HTTP 402 until you top up. This is immediate — there is no grace period and no negative balance on prepaid accounts.
• The Console stays accessible at zero balance, so you can top up, read your logs, and manage keys.
• Requests rejected for insufficient Credit are not billed.

All sales are final. Credit is not refundable, spent or unspent, and a balance cannot be paid out. The Refund & Cancellation Policy on this page sets out the detail, including how billing errors and chargebacks are handled — those are corrections and disputes, not refunds.

• Query any charge within sixty (60) days of the request that produced it, quoting the request IDs from your usage logs. After that period the charge is treated as accepted.
• If we bill you incorrectly, we will correct it by crediting your balance. If we undercharge you, we may bill the shortfall within a reasonable period.
• We do not withhold service while a good-faith dispute is investigated, provided undisputed amounts are paid.
• Card disputes and chargebacks — where you ask your bank to reverse a payment rather than raising it with us — are covered in section 7 of the Refund & Cancellation Policy.

• We may change per-model prices and account-group ratios at any time and without prior notice to you. Model pricing tracks the upstream Model Providers, whose own prices change without notice to us, so we cannot promise notice we are not given ourselves.
• The price list in the Console is authoritative and is the only place a current price is published. Check it before committing a workload, and re-check it before scaling one up. A price quoted on this website, in the documentation, or in a previous conversation is not binding.
• Price increases apply only to requests made after the change. They are never applied retroactively to usage already billed, and never to Credit you have already bought — a top-up buys a balance in Singapore dollars, and that balance is not revalued.
• Price reductions and new models may take effect immediately.
• A model may be repriced, deprecated, or withdrawn when its Model Provider does so. Check the price list in the Console before committing a workload to a specific model, and pin model versions in production.
• Enterprise pricing is fixed for the term stated in the relevant agreement.

Enterprise customers may be set up for invoiced billing under a written agreement. Where that applies, the invoice for the previous month's usage is issued at the beginning of each month and is payable within the period stated in your agreement; overdue invoices may accrue late-payment interest at the rate stated there or, absent one, the maximum permitted by law. Where an enterprise agreement conflicts with this document, the enterprise agreement prevails.

To discuss invoiced billing, committed volume, or dedicated capacity, contact hello@lulutokens.ai.

• If you close your account, or we terminate it for cause, unused Credit is forfeited. There is no payout on closure. See the Refund & Cancellation Policy.
• Fees for Services already provided remain payable.
• Credit expires after twenty-four (24) consecutive months with no successful API request and no top-up, and the account may be closed as dormant. We email the address on the account first and give you at least thirty (30) days to use the balance or top up before either happens.
• On closure, API keys stop working immediately. Export anything you need — usage logs, receipts — before you close the account.

Billing questions, invoices, and enterprise arrangements: hello@lulutokens.ai.

LuluTokens sells prepaid API credit in Singapore dollars. Credit is consumed request by request, at the per-token price published in the Console for the model you called. It is a digital service, delivered the moment your payment confirms and consumed immediately on each request.

We do not give refunds. Once Stripe confirms a top-up, that payment is final: we do not refund it, reverse it, or pay out a balance — in whole or in part, whether or not you go on to use the credit.

This holds however things turn out. It applies when the Output was not useful to you, when a Model Provider filtered or refused part of your request after processing it, when a model you relied on is repriced or withdrawn, when you change your mind, and when your own code sent requests you did not intend.

There are two reasons. Credit is delivered as soon as the payment confirms, so the thing you bought has already been supplied. And we pay Model Providers for served requests as they happen, and that cost is not recoverable for us either.

Controlling spend is your responsibility. Top up in the amounts you actually need, and keep your rate limits and client code under control.

Where a law that applies to you gives you a remedy that cannot be excluded by agreement, that right stands. Nothing here attempts to remove it.

Unused credit is not refundable either. A balance you have not spent is still credit, not money held on your behalf, and there is no route back to cash.

• It expires after twenty-four (24) consecutive months with no successful API request and no top-up. We email the address on the account before that happens and give you at least thirty (30) days to use the balance or top up.
• It cannot be transferred to another account, sold, assigned, or given away.
• It cannot be exchanged for anything other than usage of the Services.
• It is forfeited if you close your account, or if we terminate it for cause.

If we permanently discontinue prepaid Services, our sole obligation is to either let you spend the remaining balance during an announced wind-down period, or refund it where applicable law requires — see section 9 of the Terms of Service. That is the only route by which credit becomes money again.

A completed top-up is required before the API will serve any request — including requests to models listed at zero price. Free models are a price, not a free tier: they cost nothing per token, and they are available to accounts that have paid.

• Any trial, promotional, or gift credit we grant has no cash value, is never refundable, is not transferable, and may expire or be withdrawn at our discretion.
• One trial balance per person or organisation. Registering multiple accounts to collect trial credit breaches the Usage Policy and forfeits every balance involved.

• There is nothing recurring to cancel. LuluTokens is prepaid: we do not run subscriptions and we never charge your payment method on a schedule. Stopping means simply not topping up again.
• You can close your account at any time from the Console, or by asking us to close it.
• Closing the account forfeits any unused credit. There is no payout on closure.
• Closing your account revokes your API keys immediately. Export your usage logs and receipts first.

None of these are refunds, and all of them are handled on your balance rather than through your payment method.

• Requests that fail before reaching a Model Provider are never billed, so there is nothing to put right.
• A Model Provider blocking, filtering, or refusing a request is not a fault in the Services, and the request is billed where it was served and billed to us.
• If we have metered or billed you incorrectly, that is a billing error, not a refund matter — tell us and we will correct it by crediting your balance. See the Payment Terms.
• We do not offer service credits for downtime on prepaid accounts. See the Terms of Service.

Talk to us first. Email hello@lulutokens.ai before disputing a charge with your bank or card issuer. Most disputes come from an unrecognised statement line, a payment that has not yet cleared, or usage someone else on the account ran up — all of which we can usually explain or resolve the same day, and far faster than a card network can.

There are two different things people call a dispute, and they are handled differently:

• A billing dispute — you think a charge or a metered amount is wrong. Raise it with us within sixty (60) days of the request that produced it, quoting the request IDs from your usage logs. We investigate and correct genuine errors by crediting your balance. See the Payment Terms.
• A card dispute or chargeback — you ask your bank to reverse the payment. This bypasses us entirely and takes weeks to resolve through the card network.

What happens when a chargeback is filed. We may suspend the account immediately and hold the remaining balance pending the outcome, and we may recover any dispute fees charged to us. We will contest the chargeback with the evidence we hold: the payment record, the top-up record, and the metering and usage logs for the account, including timestamps, models called, token counts, and the API keys used.

A chargeback does not undo consumed credit. If the credit funded by the disputed payment has already been spent on API requests, those requests were served and paid for upstream. Where a dispute reverses money for credit you have already used, the resulting shortfall is a debt to us, and we may recover it and decline to reopen the account.

Where a dispute reverses money for credit you have not used, the corresponding credit is removed from your balance. You cannot end up holding both the money and the credit.

In short: where a payment is disputed, reversed, charged back, or found to be fraudulent, we may suspend the account immediately and deduct the corresponding Credit from its balance — including where that takes the balance to zero. Credit is not a store of value we hold for you; it is the thing the payment bought, and a payment that comes back takes the Credit with it.

If you disagree with how we handled a dispute, reply to hello@lulutokens.ai and we will review it. Nothing in this section affects your rights against your card issuer, or the dispute-resolution and governing-law terms in the Terms of Service.

We may update this policy. The version published here at the time of your payment governs that payment.

Matrix Minds Pte. Ltd. — operator of LuluTokens

• Refunds and billing: hello@lulutokens.ai
• Address: 336 SMITH STREET, #05-301, NEW BRIDGE CENTRE, SINGAPORE (050336)