Terms of Service

Nature of Service

1.1 Corporate Ownership. AI HUB is owned, operated, and maintained exclusively by Alchemist Marketing Sdn. Bhd., a private limited company incorporated under the laws of Malaysia (Company Registration No. 202101000044 (1400342-A)), with its registered corporate address at Block C-38-2 (Second Floor), Zenith Corporate Park, 1, Jalan SS 7/26, SS7, 47301 Petaling Jaya, Selangor, Malaysia.

1.2 Blind Data Conduit. The Service operates strictly as a passive third-party application programming interface ("API") aggregator and pass-through routing platform. The Service grants access to generative artificial intelligence, machine learning, and large language models ("Models") developed, hosted, and operated by independent third-party entities ("Model Providers"). Company exercises no editorial, qualitative, or manual processing control over API Inputs or Outputs.

1.3 Independent Entity Status. Company does not create, train, fine-tune, control, manage, or verify third-party Models. Company retains the right to modify, add, replace, throttle, or deprecate access to any Model or feature for legitimate operational reasons, and shall use reasonable efforts to provide notice of material changes where practicable.

1.4 Platform Features. The Service currently exposes the following functions. Company may add, change, or retire any of them at any time under Section 1.3:

• Chat, image, audio, and video generation. Models are reachable from a single creation workspace, either through your own Model Provider account ("BYOK," bring your own key) or through Company's managed credential (the "Hub Key").
• Ad Studio. Composes a still advertisement frame from a product photograph and up to two optional reference images ("cast slots"), then animates that frame into a short video. Section 2.3 governs a photograph of a real person placed in a cast slot.
• Long-form video. Assembles a multi-clip video from a script or a topic you supply, generating each clip in turn and stitching the approved takes into one finished piece.
• AI Slides. Generates a presentation deck from a prompt, an uploaded document, or both, and produces a downloadable file.
• Social publishing. Composition and publication of posts, Reels, and Stories to Facebook Pages, Instagram, Threads, TikTok, and YouTube accounts you connect, together with read-back of the engagement metrics on those posts. All five platforms are generally available: Meta approved Company's application for the Facebook Page, Instagram, and Threads permissions, and TikTok approved Company's application including direct publication to a TikTok account.
• Brand DNA. Stores a description of your brand's voice and visual identity and applies it to the text and media the Service generates for you.
• Styles and Community. Saves the words, inputs, settings and up to three kept images of a result as a reusable style. A style stays private to you until you publish it to Community, where other users can generate with it and save a copy of it. Sections 6.7 to 6.10 govern publishing.
• Hermes Managed Agents. Provisioning of autonomous AI agents that operate continuously on your instruction across messaging channels you configure. Section 7 governs Hermes.
• Developer API. Issuance of API keys ("sk-hub-…") that call Company's managed inference gateway directly from your own software. Section 4 governs the roles this creates between you, your end-users, and Company.
• Public Playground. A capped, no-account demonstration endpoint on Company's marketing site. Section 3.10 states its limits.

Eligibility, Account, and Security

2.1 Age Requirement. You must be at least eighteen (18) years of age to use the Service. By accessing or using the Service, you represent and warrant that:

• (a) you are at least eighteen (18) years of age, or have reached the age of majority in your jurisdiction of residence, whichever is higher;
• (b) you possess full civil capacity and corporate/individual authority to enter into a binding contract under applicable Malaysian law (including the Contracts Act 1950) and your home jurisdiction;
• (c) if entering on behalf of an entity, you are fully authorized to bind such entity; and
• (d) you have not been previously banned or suspended from the Service.

2.2 Sanctions and Export Controls. You warrant that you are not located in a sanctioned region or listed on restricted party lists (including OFAC SDN, EU, UN, or Malaysian sanctions lists).

2.3 Ad Studio Likeness Warranty. If you place a photograph of a real, identifiable person into an Ad Studio cast slot, you additionally warrant, strictly and without qualification, that: (a) you hold that person's affirmative permission to use their likeness in an advertisement; (b) that permission extends to the photograph being processed by third-party Model Providers located outside your jurisdiction, as described in Section 4.1; and (c) the person depicted is an adult under Section 2.1(a). Company records the moment you confirm this warranty against the generation it attaches to. Company does not verify the warranty and has no technical means to do so. A photograph submitted without satisfying all three limbs of this warranty is a breach of these Terms in itself, independent of anything generated from it.

2.4 Account Types. Individual Accounts and Organizational Accounts (administered by Admin Users). Admin Users maintain operational control over configurations, API keys, and budgets for their organization.

2.5 Account Security. You are responsible for safeguarding your account credentials and API keys and for any activities under your account, and you agree to notify us immediately of any unauthorised use of your account. An account may be opened with an email address and password, or by continuing with Google through OAuth; where Google sign-in is used, Company never receives or stores your Google password, only your name, email address, account identifier, and profile picture. Company does not currently offer two-factor authentication on password-based accounts; where you sign in with Google, your Google account's own two-factor protection governs your sign-in. Provider API keys, social platform tokens, and Hermes channel credentials are encrypted at rest using Laravel's authenticated encryption, decrypted in memory only for the duration of a request you initiated, and never returned to your browser beyond a masked tail such as sk-…4f2a. Developer API keys go further: Company stores only a one-way SHA-256 hash and a masked tail, never the plaintext, and cannot display or recover a lost key. Account passwords are stored as bcrypt hashes. These are descriptions of Company's practices, not a warranty of impenetrability.

2.6 Suspension and Termination. Company reserves the right to suspend or terminate your account or access to the Service for material breach of these Terms, non-payment, or violation of applicable law. Company shall provide reasonable notice where practicable, except where such notice would compromise security or is prohibited by law. Upon termination for any reason, all Credits remaining in your account shall be forfeited in accordance with Section 3.3. Termination stops managed spending, pauses running Hermes agents, and blocks publishing to connected social accounts; it does not, of itself, delete stored data.

2.7 Service Outages. THE SERVICE IS PROVIDED ON AN "AS IS" AND "AS AVAILABLE" BASIS. WE DO NOT WARRANT THAT THE SERVICE WILL BE UNINTERRUPTED, ERROR-FREE, SECURE, OR CONTINUOUSLY OPERATIONAL. Server outages, system latency, database corruption, or API downtime caused by third-party criminal acts—including Distributed Denial of Service (DDoS) attacks, unauthorized network hacking, malware infections, or upstream ISP failures—shall be classified as Force Majeure events. Company shall not be liable for any indirect, consequential, special, or punitive damages, or for loss of profits, revenue, or business opportunity, arising from service outages, cyber attacks, or system disruptions caused by events beyond Company's reasonable control.

2.8 Account Closure. Self-service account closure is not yet available. To close an account, email [email protected] from the address on the account; Company confirms your identity and closes the account by hand. Any remaining Credit balance shall be forfeited in accordance with Section 3.3. Sections 7, 8, 9, 10, and 11 survive termination or closure, along with any indemnity you have given Company.

Credits and Payment

3.1 Pre-Paid Credits. Access to Models requires pre-purchasing non-transferable, non-assignable digital credits ("Credits"). Credits represent solely a limited, revocable license to route API calls. Credits carry no monetary value, do not constitute personal property, and cannot be redeemed, exchanged, or transferred.

3.2 Payment Processing. All payments are processed by Stripe, a PCI-DSS-compliant payment processor. We do not store your full payment card details. By making a payment, you agree to Stripe's terms of service and privacy policy. The Credit balance is denominated in United States dollars and is prepaid: there is no invoice, no credit line, and no charge to a card without an affirmative top-up request. The minimum top-up is five (5) US dollars and the maximum is ten thousand (10,000) US dollars per transaction. Card payments settle in US dollars; FPX online banking and Alipay do not support a US-dollar charge on Company's Malaysian Stripe account, so a top-up by either rail is charged to you in Malaysian ringgit at the rate shown on the confirmation screen before you commit, and your Credit balance is credited the full US-dollar amount you requested regardless of that conversion. Where you draw on the Hub Key, the Credit deduction includes Company's margin over the Model Provider's own charge, up to twenty percent (20%); where you connect your own Model Provider account ("BYOK"), that provider bills you directly at its own rates and Company adds no markup.

3.3 Credit Purchases Are Final. Every purchase of Credits is strictly final and non-refundable. "Non-refundable" means that no refund, reimbursement, or repayment of any Credits (whether used or unused) shall be made upon termination of your account, except as required by applicable law or as Section 3.3(a) below describes. All Credits remaining in your account at the time of termination shall be forfeited.

3.3(a) Voluntary Exception: the Hermes Plan-Fee Refund. Of every charge the Service can generate, Company chooses to make exactly one kind capable of being refunded on request: a Hermes managed agent plan-fee charge, and only in the manner this Section 3.3(a) describes. Model or token spend, Ad Studio generations, Developer API usage, and every other charge of any kind remain non-refundable under Section 3.3 without exception.

• (i) By Request, Not Automatic. You must submit a written refund request to Company within thirty (30) calendar days of the plan-fee charge you are disputing. A request submitted after that window is time-barred.
• (ii) Company's Discretion. Company reviews each request on its own facts and may approve or reject it in its sole and absolute discretion.
• (iii) Credit to the Wallet Only. An approved refund is credited to your Credit balance and nothing else. It is never paid to the original payment method, card, bank account, or payment instrument, and never issued as cash.
• (iv) One-Hundred-Eighty-Day Cooldown. Once a refund request is accepted, your customer account may not submit another refund request for one hundred eighty (180) days from the date of acceptance. The cooldown runs against your customer account as a whole and is triggered only by an accepted request.
• (v) Waiver Is Discretionary and Creates No Precedent. Company may, in its sole discretion, waive or shorten the cooldown for any account. Any such waiver is a one-time accommodation on the facts of that account alone and creates no right to a further waiver.
• (vi) Deletion Precedes a Request. A refund request may be submitted only for a Hermes agent that has already been deleted. Deletion is permanent under Section 7.2 and is not reversed or made conditional by a request's submission, pendency, approval, or rejection.

3.4 Anti-Money Laundering (AML). You warrant that all funds originate from lawful sources and do not violate the Malaysian AMLA 2001 or international AML standards. Company reserves the right to request Know-Your-Customer (KYC) details if suspicious transaction patterns occur.

3.5 API Consumption and Upstream Errors. Credits are deducted automatically based on API usage metrics logged by Company's servers. Company offers no refunds or credits for third-party Model hallucinations, latency, timeouts, deprecations, or formatting errors. Usage Data and transaction receipts are retained for seven (7) years under Section 82 of the Malaysian Income Tax Act 1967 and the AMLA 2001, overriding any account deletion request.

3.6 Forfeiture on Breach. Account suspension due to a breach of these Terms (including prompt injection, scraping, rate-limit bypassing, or chargeback threats) results in the immediate forfeiture of all unused Credits to Company as liquidated damages without right of recovery.

3.7 Zero-Balance Handling and the Hermes Plan-Fee Grace Period. Managed spending on the Hub Key and Developer API stops at a zero Credit balance; there is no overdraft, and a running Hermes agent whose owner's balance reaches zero is paused rather than allowed to run into deficit. The Hermes flat plan fee is treated more leniently: if a renewal falls due and the balance will not cover it, the agent keeps running and Company sends reminder notices, and the agent is stopped only fourteen (14) days after an unpaid renewal if the balance is still insufficient. Its configuration and data are retained, not deprovisioned, so restoring the balance and redeploying is available after a stop.

3.8 Pricing. Real-time pricing displayed on the Site or returned in API headers at the execution timestamp shall govern Credit deduction rates. Company reserves the right to correct manifest pricing errors that result from system malfunctions.

3.9 Chargebacks. Initiating an unauthorized chargeback or payment reversal constitutes a material breach. Company may terminate access, forfeit Credits, and pursue legal recovery for disputed amounts and associated costs.

3.10 Public Playground Limits. The Playground is a capped, no-account demonstration of a live Model, and using it constitutes acceptance of these Terms even without an account. Each visitor is limited to four (4) requests per minute and twenty (20) requests per day, counted by IP address; a prompt is capped at five hundred (500) characters and a reply at two hundred fifty-six (256) tokens; and a daily budget applies across all visitors combined, beyond which the Playground stops answering until the following day. Section 5(i) governs attempts to circumvent these limits.

Third-Party Services and Model Providers

4.1 Third-Party Model Providers. The Service provides access to Models operated by third-party Model Providers. Company does not control, and is not responsible for, the operation, availability, or output of such Models.

4.2 No Warranty as to Output. Company does not warrant the accuracy, reliability, or completeness of any Output generated by third-party Models. You are solely responsible for verifying the Output before relying on it.

4.3 Third-Party Terms and Routing. Your use of third-party Models may be subject to the terms and conditions of the respective Model Providers, and you are responsible for reviewing and complying with them. On BYOK, the relationship is direct: you hold the account with the Model Provider and Company never sees your invoice from that provider. On the Hub Key path, Company holds the account, and your request is routed through an inference aggregator (currently MyTokenGate or ATP, depending on the Model) and from there to the underlying Model Provider, which may include OpenAI, Anthropic, Google, DeepSeek, MiniMax, ByteDance, and others.

4.4 Changes to Third-Party Services. Company may, at any time, add, modify, or remove access to any third-party Model or service. Company shall use reasonable efforts to provide notice of material changes, but is not liable for any loss or damage arising from such changes.

4.5 The Developer API: Roles and Flow-Down. Where you mint a Developer API key and use it to serve your own users, you are the controller of those users' personal data and Company is the processor, acting solely on your documented instruction to return the completion you requested. You must bind your own users to terms at least as protective as these, including Section 5, and you are liable to Company for everything done through your key exactly as if you had done it yourself. A generic Data Processing Agreement is published at aimodelhub.ai/dpa; if your compliance regime requires a signed or customer-specific processor agreement, email [email protected]. Each key is limited to one hundred twenty (120) requests per minute by default; requests beyond that receive an HTTP 429, and Company may throttle, suspend, or revoke a key that sustains abnormal volume or threatens the stability of the gateway.

Prohibited Uses

You agree not to use the Service:

• (a) to generate, upload, or transmit any content that is unlawful, harmful, threatening, abusive, harassing, defamatory, vulgar, obscene, libellous, invasive of another's privacy, hateful, or racially, ethnically, or otherwise objectionable;
• (b) to impersonate any person or entity;
• (c) to violate any applicable law or regulation;
• (d) to infringe upon the intellectual property rights of any third party;
• (e) to transmit any viruses, worms, Trojan horses, or other destructive items;
• (f) to interfere with or disrupt the Service or servers or networks;
• (g) to attempt to gain unauthorised access to any part of the Service, or to reach another customer's data, container, or Hermes agent;
• (h) to develop or train competing AI models or services using Outputs obtained through the Service;
• (i) to bypass rate limits, geoblocking, or security controls, or to automate the Public Playground, rotate IP addresses to defeat its limits under Section 3.10, or otherwise use it as a free inference endpoint in place of an account;
• (j) to conduct unauthorized red teaming, prompt injection attacks, or jailbreaking;
• (k) to resell, sublicense, or share a Developer API key, or use one to operate a competing inference gateway, proxy, or reseller service;
• (l) to use the Service in a way that breaches the terms of any Model Provider or connected social platform in the request path, as described in Section 4.3; or
• (m) to publish a style that contains a third party's brand, characters or artwork you have no right to share, or a kept image of a real person, or to report a style you know does not breach these Terms; or
• (n) to submit bad-faith legal reports against Company.

5.1 Ad Studio and Hermes Attribution. Uploading a photograph without satisfying the warranty in Section 2.3 is a breach of this Section in itself. Your Hermes agents act on your instruction and are your responsibility; an agent that breaches this Section breaches it on your behalf, and the fact that you did not personally issue the instruction that produced the output is not a defense.

5.2 Company may investigate suspected breaches of this Section and cooperate with law enforcement and with the Model Providers and platforms in the request path. Section 2.6 governs what Company may do about a breach.

User Content and Intellectual Property

6.1 User Content. You retain all ownership rights in any content ("User Content") that you submit, post, or display on or through the Service.

6.2 License to Company. By submitting User Content to the Service, you grant Company a worldwide, royalty-free license to use, reproduce, store, and transmit your User Content for the duration necessary to provide the Service to you: routing a prompt to the Model you chose, holding a generated asset in your media library, publishing a post to a social account you connected, and displaying any of it back to you. This license ends when you delete the relevant User Content from the Service, or when your account is terminated. Company does not use it for its own marketing and does not sublicense it beyond what this Section and Section 4 describe.

6.3 License to Model Providers. You acknowledge that when you submit User Content to the Service, such content is transmitted to third-party Model Providers. The processing of your User Content by such Model Providers is governed by their respective terms of service.

6.4 Output Ownership. Subject to the terms of the applicable Model Provider, you own all rights to the Outputs generated by the Service from your Inputs. Company does not claim any ownership rights in Outputs. In several jurisdictions, including the United States, purely machine-generated output may not attract copyright protection at all, and the position remains unsettled elsewhere; Company gives no warranty of exclusivity or copyrightability in any Output, and two users submitting similar prompts to the same Model may receive similar or identical Output with no claim against each other or against Company arising from that similarity.

6.5 No Training. Company does not use your User Content or Outputs to train its own machine learning models.

6.6 Publishing License Pass-Through. When you publish through the Service to a connected social account, you grant that platform whatever license its own terms require to display the post. That grant runs between you and the platform directly; Company is the mechanism, not a party to it.

6.7 Publishing a Style to Community. A style you publish to Community is User Content you choose to share with other users of the Service. Publishing is available to an account that has funded its wallet at least once. By publishing a style, you grant: (a) each other user of the Service a worldwide, royalty-free, non-exclusive licence to view the style, generate Outputs with it through the Service, save a copy of it to their own account, and keep using that copy under Section 6.10; and (b) Company a worldwide, royalty-free, non-exclusive licence to host, store, reproduce, transmit and publicly display on the Service the style's name, tagline, words, cover image and kept images, and to have automated systems and Company staff review them, to the extent Company needs to operate Community, check the style against Section 5 and the Community rules, and act on reports under Section 6.9. The licence in (b) does not permit Company to use your style in its own marketing without your permission. Section 6.4 gives each user ownership of the Outputs that user generates with your style.

6.8 Your Public Name. Company shows a public name ("handle") beside each style you publish, and does not display your real name or email address. A handle is 3 to 30 characters of lowercase letters, numbers, dots and underscores. Company treats two handles that differ only by dots or underscores as the same handle, and you may not choose a handle Company reserves for AI HUB, its staff or Model Providers. Company assigns you a handle when you create your account, and you may change it on your account page once every thirty (30) days. When you change it, Company retires your previous handle and does not make it available to any other account. You may not choose a handle that impersonates a person, brand or organisation, and Company may withdraw such a handle under Section 6.9.

6.9 Checks, Reports, Takedown and Appeal. (a) Automated check. Company runs an automated check on each publication, and on each edit to a published style, before other users can see it. The check applies Section 5 and the Community rules, including the rule that a kept image shows no person. If a style fails the check, or Company cannot run the check at that moment, Company does not publish it; the version already published, if any, stays live until an edit passes. (b) Reports. Any user may report a published style for a reason the Service lists. When several distinct, established accounts report the same style, Company hides it from Community until Company reviews the reports; while it is hidden, only you can use it. (c) Takedown. Company may hide, unpublish or remove a published style, withdraw a handle, or stop copies of a removed style from generating, where Company reasonably considers that the style breaches these Terms, infringes a third party's rights, or is the subject of a valid legal notice. If Company removes a style under Section 5 for sexual content, a real person's likeness or a third party's rights, Company may also stop saved copies of that style from generating. (d) Appeal. If Company refuses, hides or removes your style, you may edit it and resubmit it through the check, make it private, or ask Company to review the decision. Company notifies you by email when it hides or removes a published style, and you have thirty (30) days from that email to ask for a review, quoting the style's name and your handle. Where Company refused the style at publication rather than removing it, the same thirty (30) days run from the refusal. Company will review the decision in good faith and reply within a reasonable time. Company's decision after that review is final for the purposes of the Service and does not affect your rights under Section 11. To report a style that infringes your rights without holding an account, email [email protected] with the style's name, the handle and the right you hold.

6.10 Unpublishing, Copies, and Account Closure. You may unpublish a style at any time, and Company removes it from Community at once. A copy another user saved before you unpublished is a frozen snapshot of the style at the moment that user saved it, including its words and kept images. That copy stays in the other user's account and keeps working after you unpublish, edit or delete your style, or close your account, unless Company removes it under Section 6.9(c). Unpublishing does not affect Outputs anyone already generated with your style. When Company closes your account, Company removes your published styles from Community.

Hermes Managed Agents

7.1 Agent Configuration and Behaviour. You are responsible for the configuration, instructions, and behaviour of any Hermes agent you provision. Company does not control or monitor the behaviour of Hermes agents beyond providing the underlying infrastructure. You are solely responsible for the actions and outputs of your agents.

7.2 Agent Data and Isolation. Each agent's data, including its conversation history and any files it handles, is deleted when you delete the agent; deletion destroys its container, memory, uploaded documents, and channel connections irreversibly and is unaffected by whether a refund request under Section 3.3(a) has been submitted, is pending, or has been approved or rejected. Each Hermes agent runs in its own container with Linux capabilities dropped, privilege escalation blocked, and bounded CPU, RAM, and storage. Default-deny network egress restricts an agent to the AI Model Provider and messaging channel you configured for it. No inbound network path exists into a running agent or its container, and an agent can never reach another customer's agent, container, or data. This architecture is a description of present engineering practice, not a warranty that it cannot be defeated.

7.3 Development Status. Hermes is an evolving platform Company is still building and hardening. Its features and safeguards change between releases, and Company does not warrant that it is stable, secure against every threat, or free of defects, notwithstanding the isolation architecture in Section 7.2. Company ships security patches as new Hermes runtime versions; enabling auto-update in the agent creation wizard keeps an agent on the newest vetted version automatically, while pinning a specific version makes you solely responsible for reviewing and updating it, and running an outdated pinned version is at your own risk.

User Representations and Indemnification

8.1 By using the Service, you represent and warrant that:

• (a) you are the creator and owner of, or have the necessary licenses, rights, consents, and permissions to use, all Inputs you submit to the Service;
• (b) your Inputs do not and will not infringe, violate, or misappropriate any third-party right;
• (c) your Inputs do not and will not slander, defame, libel, or invade the right of privacy of any other person;
• (d) your use of the Service does not and will not cause us to violate any law or regulation.

8.2 You agree to indemnify and hold harmless Alchemist Marketing Sdn. Bhd., its directors, officers, employees, and agents from and against any and all claims, damages, losses, liabilities, costs, and expenses (including reasonable legal fees) arising out of or relating to:

• (a) your Inputs, User Content, or published materials;
• (b) your violation of any term of this Agreement;
• (c) your violation of any applicable law or regulation, including but not limited to data protection laws;
• (d) any dispute between you and a third party arising from your use of the Service.

8.3 Named Triggers, Without Limitation. Without limiting the generality of Section 8.2, the indemnity extends specifically to: a claim, demand, or proceeding arising from a photograph you uploaded to an Ad Studio cast slot under Section 2.3; content generated, sent, or acted on by a Hermes agent operating on your instruction under Section 7.1; and a claim brought by your own end-users, or by a regulator on their behalf, arising from your use of the Developer API under Section 4.5. The indemnity also extends to a claim arising from a style you published to Community under Section 6.7, including a claim by a user who generated with it or saved a copy of it.

Disclaimer of Warranties

9.1 "AS IS" and "AS AVAILABLE" Service. THE SERVICE AND OUTPUTS ARE PROVIDED "AS IS" AND "AS AVAILABLE". TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, WE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE, AND NON-INFRINGEMENT. WE DO NOT WARRANT THAT THE SERVICE WILL BE UNINTERRUPTED, ERROR-FREE, SECURE, OR CONTINUOUSLY OPERATIONAL.

9.2 AI Output Reliability. Output may be inaccurate, out of date, biased, offensive, or fabricated. Reviewing Output before relying on it, publishing it, sending it to a customer, or acting on it is your responsibility alone. Nothing an AI Model produces through the Service is legal, medical, financial, or professional advice.

9.3 Playground Demonstration Disclaimer. The Public Playground under Section 3.10 is a demonstration and carries no warranty of any kind.

Some jurisdictions do not allow the exclusion of certain warranties. Where that is the case, the exclusions above apply only to the extent permitted, and your mandatory statutory rights are unaffected.

Limitation of Liability

10.1 Exclusion of Indirect Damages. TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, IN NO EVENT SHALL WE BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, OR PUNITIVE DAMAGES, OR FOR ANY LOSS OF PROFITS, REVENUE, DATA, USE, GOODWILL, OR OTHER INTANGIBLE LOSSES, ARISING OUT OF OR RELATING TO THIS AGREEMENT OR THE SERVICE, WHETHER BASED ON CONTRACT, TORT (INCLUDING NEGLIGENCE), OR OTHERWISE, EVEN IF WE HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

10.2 Limitation of Liability. TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, OUR AGGREGATE TOTAL LIABILITY FOR ANY AND ALL CLAIMS ARISING OUT OF OR RELATING TO THIS AGREEMENT OR THE SERVICE SHALL BE LIMITED TO THE GREATER OF: (A) THE TOTAL AMOUNT PAID BY YOU FOR THE SERVICE IN THE TWELVE (12) MONTHS IMMEDIATELY PRECEDING THE EVENT GIVING RISE TO THE CLAIM; OR (B) ONE HUNDRED UNITED STATES DOLLARS (USD 100.00). This clause is a limitation, not an absolute exclusion. Under Section 29 of the Contracts Act 1950, clauses which absolutely restrict a party from enforcing their rights through legal proceedings are void. This clause does not prevent you from bringing a claim—it merely limits the recoverable damages.

10.3 Exceptions. THE LIMITATIONS IN SECTIONS 10.1 AND 10.2 SHALL APPLY TO THE FULLEST EXTENT PERMITTED BY APPLICABLE LAW.

10.4 Force Majeure. We shall not be liable for any delay or failure to perform our obligations where such delay or failure results from any cause beyond our reasonable control, including but not limited to: acts of God, war, terrorism, pandemic, government actions, network failures, cyber attacks, or third-party provider outages.

Governing Law and Dispute Resolution

11.1 Governing Law. This Agreement shall be governed exclusively by the laws of Malaysia.

11.2 Mandatory Arbitration. ANY DISPUTE, CLAIM, OR CONTROVERSY ARISING OUT OF OR RELATING TO THIS AGREEMENT OR THE SERVICE SHALL BE RESOLVED BY BINDING ARBITRATION ADMINISTERED BY THE ASIAN INTERNATIONAL ARBITRATION CENTRE (AIAC) IN KUALA LUMPUR, MALAYSIA, IN ACCORDANCE WITH ITS RULES THEN IN EFFECT. YOU AGREE TO GIVE UP YOUR RIGHT TO GO TO COURT to assert or defend your rights under this Agreement, except for matters that may be taken to the Small Claims Court of Malaysia. YOU AGREE TO GIVE UP YOUR RIGHT TO PARTICIPATE IN A CLASS ACTION or class arbitration against us.

11.3 Small Claims Exception. Either party may bring proceedings in the Small Claims Court of Malaysia for claims within that court's jurisdictional limit, notwithstanding the arbitration clause above.

Modifications to These Terms

12.1 Company reserves the right to modify these Terms. Company will provide notice of material changes at least thirty (30) days in advance via email or through the Service. If you do not agree to the modified Terms, you may terminate your account without penalty within that period: Company will not charge any early-termination fee, cancellation fee, or penalty as a result, though this does not entitle you to a refund of any Credits previously purchased except as Section 3.3 already allows. Your continued use of the Service after the notice period constitutes your acceptance of the modified Terms.

12.2 Non-material Changes. Company may make non-material changes to these Terms at any time without advance notice. Typographical corrections, broken links, and formatting fixes are non-material; anything that changes what you are agreeing to is material and follows Section 12.1.

12.3 How Acceptance Is Recorded. Beyond the notice in Section 12.1, acceptance of these Terms and the Privacy Policy is separately captured when you register, sign in with Google for the first time, and again at each account-critical action, including adding funds, connecting a social account, storing a Model Provider key, minting a Developer API key, and deploying a Hermes agent. Each acceptance is recorded against your user ID, the action being taken, the moment of acceptance in Coordinated Universal Time, the version of each document as it then stood, a SHA-256 digest of the exact text of those versions, the display language, and the IP address and browser the acceptance came from. This acceptance gate operates in the console session and does not intercept a request authenticated with a Developer API key minted before a new version was published, nor a Hermes agent already running; both continue operating on a superseded version until you separately act, and Company will give notice at the email address on your account before deliberately disabling a key or agent for non-acceptance.

General Provisions

13.1 Entire Agreement. These Terms, together with the Privacy Policy and the Data Processing Agreement, constitute the entire agreement between you and Company regarding the Service.

13.2 Severability. If any provision of these Terms is held to be invalid or unenforceable, the remaining provisions shall continue in full force and effect.

13.3 Waiver. The failure of Company to enforce any right or provision of these Terms shall not constitute a waiver of such right or provision.

13.4 Assignment. You may not assign these Terms without the prior written consent of Company. Company may assign these Terms without your consent.

13.5 No Agency. Nothing in these Terms shall be construed as creating an agency, partnership, joint venture, or employment relationship between you and Company.

13.6 Version Integrity. The published text of each version of these Terms is fingerprinted at the moment of publication, and a change to that text which is not consciously recorded against the manifest fails Company's build. Superseded versions are not deleted; email [email protected] with a version number for its exact text.

Contact Us

Alchemist Marketing Sdn. Bhd. (Company Registration No. 202101000044 (1400342-A))
Block C-38-2 (Second Floor), Zenith Corporate Park,
1, Jalan SS 7/26, SS7, 47301 Petaling Jaya, Selangor, Malaysia.

Email: [email protected]