LIQUID INFERENCE TERMS OF SERVICE

Effective Date: November 4, 2026

1. AGREEMENT AND ELIGIBILITY

1.1 These Terms of Service (“Terms”) are a legally binding agreement between Architect Financial Technologies Inc., a Delaware corporation (“Architect,” “we,” or “us”) and the person or organization that accepts them (“you”). You accept these Terms by creating an account, clicking to accept, or using the Services. If you accept for an organization, you confirm that you are authorized to bind it, and “you” means that organization. SECTION 19 REQUIRES MOST DISPUTES TO BE RESOLVED BY BINDING INDIVIDUAL ARBITRATION AND WAIVES JURY TRIALS AND CLASS ACTIONS.

1.2 You must be at least 18 years old, or the age of majority where you live if higher, and able to enter into a binding contract. You may not use the Services if you are located, resident, or organized in, or owned or controlled by persons located in, a country or region subject to comprehensive territorial U.S. sanctions; if you are on a U.S. government restricted-party or sanctions list; or if Architect has previously terminated your access for violating these Terms.

1.3 Architect may refuse registration, require identity or business verification, or restrict account features where it reasonably considers this necessary for security, fraud prevention, sanctions screening, legal compliance, or abuse prevention.

2. DEFINITIONS

2.1 “Services” means Liquid Inference, including the websites at https://inference.ai.exchange, the APIs, dashboards, documentation, and related support.

2.2 “Provider” means an independent third party approved by Architect to quote prices for and serve Requests on Liquid Inference.

2.3 “Request” means an inference request you submit through the Services. “Inputs” means the prompts, instructions, images, files, and other content in a Request. “Outputs” means the content returned in response to a Request. Inputs and Outputs together are “Customer Content.”

2.4 “Routing Rules” means the requirements you set for a Request, such as model, Provider, region, Data Policy Designation, speed, price cap, and routing policy, together with any defaults that apply when you do not set them.

2.5 “Quote” means the price per unit (for example, per token) that a Provider offers for serving a model at a given time.

2.6 “Maximum Charge” means the highest amount a Request can cost, as shown or returned by the Services before the Request is processed.

2.7 “Model Terms” means the licenses and usage policies that a model’s developer applies to that model.

2.8 “Data Policy Designation” means the label shown for each Provider’s models (ZDR, Limited Retention, or Training), as described in Section 5.2.

3. THE SERVICES

3.1 Quotes and routing. Providers set and change their Quotes at their discretion. When you submit a Request, the Services identify the Providers whose offerings satisfy your Routing Rules. If you specify a particular Provider and model, we route the Request to it if it is available. Otherwise, Architect selects among eligible Providers in its discretion, based on factors that may include price, performance, test results, capacity, geography, and the routing policy you select. Architect may also stop or limit routing to any Provider at any time.

3.2 Maximum Charge. You will not be charged more than a Request’s Maximum Charge. You pay only for units actually produced, at the matched Quote, plus Architect’s applicable service fees.

3.3 Failed Requests. If the matched Provider fails to complete a Request, the Services may retry it with the next eligible Provider, still subject to your Routing Rules and the Maximum Charge. You are not charged for a Request that fails.

3.4 No eligible Provider. If no Provider can serve a Request, it fails and you are not charged. Stricter Routing Rules may reduce the number of eligible Providers, raise prices, or cause Requests to fail. Setting Routing Rules appropriately is your responsibility.

3.5 Location. Region settings rely on the processing countries each Provider discloses. We may pass your Request’s country of origin to Providers that require it. Location data may be inaccurate, and Architect is not responsible for inaccuracies it could not reasonably detect.

3.6 Architect’s role. Architect operates a marketplace. Inference for each Request is provided to you by the Provider that serves it, not by Architect. Architect provides the exchange, routing, billing, and related tools, and is not responsible for a Provider’s performance except as these Terms expressly state.

4. ACCOUNTS AND API KEYS

4.1 API Keys. Keep your account information accurate and your passwords and API keys confidential. Do not share API keys publicly, embed them in client-side code, or publish them in code repositories. You are responsible for all activity and charges under your account and API keys, whether or not you authorized them, except to the extent caused by Architect’s breach of these Terms. Until you notify us of a suspected compromise, we may treat all activity through your account or API keys as authorized by you. Notify us promptly at [email protected] or through in-app support if you suspect a compromise. We may disable, rotate, or restrict account credentials or keys to protect you or the Services.

4.2 Organization accounts. If Architect offers organization accounts, administrators control an organization account’s settings, and members may use the Services only as those settings allow. The organization is responsible for its members’ use.

4.3 Limits. Architect may apply rate limits, spend caps, concurrency limits, safety filters, and other technical restrictions to protect the Services, manage capacity, prevent abuse, or meet Provider requirements.

5. CUSTOMER CONTENT

5.1 As between the parties, you retain your rights in Customer Content, and Architect claims no ownership of Outputs. Your rights in Outputs may also be subject to Model Terms. You grant Architect, and the Provider serving each Request, a non-exclusive, worldwide license to host, transmit, copy, and process Customer Content as needed to provide, secure, and support the Services and to comply with law.

5.2 Data handling. Architect does not store the content of Inputs or Outputs; it processes them only in transit and in memory to route Requests and return Outputs. Each Provider’s models carry a Data Policy Designation: (a) ZDR, where the Provider has agreed not to retain Customer Content after completing the Request, other than short-lived in-memory caching and automated safety screening; (b) Limited Retention, where the Provider may retain Customer Content for the period and purposes shown in its listing, such as abuse prevention, security, and legal compliance; or (c) Training, where the Provider may use Customer Content to train or improve its models (Section 5.9). You can set your Routing Rules to use only ZDR models. Architect requires Providers, by contract, to follow their designations and not to use Customer Content for model training except for Training models. Architect relies on Providers’ disclosures and attestations, does not control their systems, and makes no promise about their practices except as these Terms and the Data Processing Agreement state.

5.3 You confirm that you have all rights, consents, and notices needed for Customer Content and its processing under these Terms.

5.4 Unless Architect has expressly agreed in a signed writing (which it can do only where the Providers serving your Requests also permit it), you must not (a) submit any Restricted Data through the Services, or (b) request that Providers provide to you any such Restricted Data (except for Restricted Data related to you and obtained by Providers from public sources). “Restricted Data” means:

(a) health information, including protected health information and medical records;

(b) payment card data, including card numbers, security codes, and magnetic-stripe or chip data;

(c) bank, brokerage, or other financial account numbers, online banking credentials, or other information that gives access to financial accounts, or other regulated financial information;

(d) passwords, API keys, access tokens, private keys, seed phrases, or other authentication secrets, whether yours or anyone else’s;

(e) government identification numbers or documents, such as Social Security, passport, driver’s license, or tax identification numbers;

(f) biometric data, such as fingerprints, face or voice templates, or retina scans;

(g) personal information about children under 18;

(h) sensitive personal information or special categories of personal information under applicable privacy laws, such as information about racial or ethnic origin, religious beliefs, sexual orientation, precise geolocation, or criminal history;

(i) material non-public information about any issuer of securities, or other information you are prohibited by law or a duty of confidentiality from sharing;

(j) export-controlled technical data or classified information;

(k) data that a law, regulation, court order, or contract prohibits you from sharing with third-party processors;

(l) child sexual abuse material, malware, or other content that is illegal to possess or distribute; or

(m) any other data subject to special legal or regulatory restrictions (including, without limitation, “human ‘omic data,” “biometric identifiers,” “precise geolocation data,” “personal health data,” “personal financial data,” “government-related data,” “covered personal data” or “combined data” under 28 CFR Part 202).

Requests are sent to independent Providers, so Architect cannot control how Restricted Data would be handled if submitted. Architect may block, delete, or report Requests it reasonably believes contain Restricted Data, and you are responsible for the consequences of submitting it.

5.5 Architect may use automated tools to screen Inputs for inappropriate material and malicious code, and may block Requests that appear to violate these Terms. Architect will report apparent inappropriate material as required by law. Architect has no obligation to monitor Customer Content.

5.6 Your end users. If you make the Services available to users of your own products, you are responsible for their Requests and for their compliance with these Terms and Model Terms.

5.7 Content standards. You must not submit Customer Content that infringes, misappropriates, or violates any intellectual property, privacy, publicity, or confidentiality right, contractual restriction, or law.

5.8 Aggregated data. Architect may collect and use aggregated or de-identified technical data, such as request volumes, latency, error rates, and security signals, to operate, secure, and improve the Services, as long as it does not identify you or any individual. Architect also keeps account, billing, and usage metadata about each Request, such as timestamps, token counts, request identifiers, and country of origin, but not its content. Architect will not use Customer Content to train AI models.

5.9 Training models. Some Providers offer models with a Training designation. Architect routes your Requests to them only if you turn on that option in your Routing Rules. If you do: (a) you must not include personal information in those Requests; (b) the Provider may use the Customer Content in them to train or improve its models under its own terms; (c) that content may not be capable of deletion or return; and (d) Architect does not warrant the Provider’s training practices. Turning the option off applies to future Requests only.

6. OUTPUTS

Outputs are generated by AI models and may be inaccurate, incomplete, offensive, or similar to Outputs given to others. Different Providers serving the same model may return different Outputs. You are responsible for evaluating Outputs, and for qualified human review before using them. Architect does not provide legal, medical, financial, investment, or other professional advice and does not warrant that output is accurate, suitable, protectable, or free of third-party rights.

7. PROVIDERS AND MODEL TERMS

7.1 Providers are independent third parties, not Architect’s employees or agents. Architect reviews Provider deployments before listing them and tests Providers periodically, but does not guarantee any Provider’s performance, security, or compliance. Architect may remove or limit any Provider at any time. For data protection purposes, Providers act as Architect’s subprocessors as described in the Data Processing Agreement between such Provider and Architect.

7.2 By requesting a model, you agree to its Model Terms, as updated by its developer, and you are responsible for reviewing them before use. You must not use VPNs, proxies, or other means to reach a model that is restricted to you, and you will provide information Architect reasonably requests to confirm your compliance with Model Terms.

7.3 We may suspend or limit your access to a model if we reasonably believe you violated its Model Terms, or if its developer or a Provider requires it, with advance notice where practical. Where Model Terms or a Provider’s usage policy are stricter than these Terms, the stricter terms govern Requests to that model or Provider.

7.4 Provider terms. You must not use the Services in a way that would cause Architect or any Provider to breach the terms or usage policies of a model developer or other upstream provider. Providers may impose their own content restrictions, technical limits, and data-processing terms, which apply to Requests they serve. As between you and Architect, these Terms govern; Model Terms and Provider terms govern your use of the relevant model or Provider to the extent the model developer or Provider requires.

8. FEES AND PAYMENT

8.1 Billing. You authorize Architect and its payment processor, Stripe, to charge your selected payment method for the amounts you choose. Payments are added to your account as prepaid credits (“Credits”), which are used to pay for Requests and Architect’s fees.

8.2 Charges. Each Request is charged at the matched Quote for units produced, plus Architect’s fee, Itemized charges are available in your account. If we change Architect’s fee, we will give you advance notice, and the change will apply only to Requests made after it takes effect.

8.3 Refunds. Credit purchases are non-refundable, except where Architect confirms a billing error, as stated in Section 15.3, or as required by law. Architect’s fees are not refundable.

8.4 Credits. Credits are not legal tender or currency, have no equivalent value in fiat currency, do not act as a substitute for fiat currency, and do not constitute or confer any personal property right. Credits are not redeemable, refundable, or exchangeable for any sum of money or monetary value, except for (i) the refunds described in Section 8.3, (ii) the refund of remaining Credits on termination described in Section 15.3, or (iii) as required by applicable law. Credits are non-transferable, including between accounts, and may be used only in connection with the Services. Credits are not deposits or stored-value accounts and earn no interest for you. Architect does not provide banking, deposit-taking, money transmission, stored-value, securities, investment advisory, brokerage, or custody services through the Services.

8.5 Disputes and taxes. Notify us of a disputed charge within 30 days. Charges exclude taxes; you are responsible for applicable sales, use, value-added, and similar taxes.

8.6 Adjustments and payment issues. Architect may correct, deduct, or reinstate balances to address billing or technical errors, fraud, chargebacks, unauthorized payments, or misuse. Architect may suspend or restrict access if a payment fails, payment credentials are invalid, a chargeback is initiated, or fraudulent payment activity is suspected.

9. ACCEPTABLE USE

You must not, and must not help others or cause us to:

(a) violate any law, sanctions program, export control, or third-party right;

(b) create or distribute malware, or engage in unauthorized access, credential theft, fraud, unlawful surveillance, or other unlawful harm;

(c) generate child sexual abuse material or any content that sexualizes minors;

(d) probe, disrupt, or overload the Services, or circumvent rate limits, Routing Rules, Maximum Charges, billing, or security controls;

(e) manipulate the exchange, including by submitting Requests designed to distort Quotes, rankings, or routing;

(f) resell or share access to the Services or API keys, except as Architect permits in writing and the applicable Provider’s terms allow;

(g) scrape the Services or extract non-public data, except through documented APIs within applicable limits;

(h) present Outputs as human-generated where disclosure is required, or use Outputs to impersonate others or deceive;

(i) create multiple accounts or misrepresent your identity to get around limits, promotional-credit restrictions, or a suspension;

(j) bypass a model’s safety controls (for example, by jailbreaking);

(k) fund your account with the proceeds of crime, or use the Services to launder funds;

(l) generate, distribute, or facilitate content that is unlawful, defamatory, harassing, abusive, deceptive, fraudulent, or exploitative;

(m) generate instructions for, or otherwise facilitate, violence, terrorism, violent extremism, self-harm, or the development or use of weapons, including chemical, biological, radiological, or nuclear weapons;

(n) send spam or phishing messages, or create botnets, ransomware, spyware, or credential-harvesting tools;

(o) conduct card testing, use stolen or unauthorized payment credentials, or otherwise attempt to process fraudulent payments through the Services or Architect’s payment processors;

(p) conduct unlawful biometric identification or categorization, or unauthorized profiling or tracking of individuals;

(q) use the Services or Outputs as the sole basis for decisions that significantly affect a person’s legal rights, finances, credit, employment, housing, education, insurance, healthcare, or access to essential services, without appropriate human review and any legally required safeguards;

(r) run bots, load tests, automated prompt-injection pipelines, or other automated systems in a way that degrades the Services or other users’ access;

(s) extract or attempt to derive models, model weights, system prompts, routing logic, or non-public features of the Services, or publish misleading benchmarks of the Services or any Provider;

(t) use the Services in a way that causes Architect, a Provider, or a model developer to breach its terms, usage policies, or applicable law;

(u) violate anti-money-laundering, anti-bribery, data protection, consumer protection, or export control laws;

(v) access the Services from a location, or on behalf of a person, that a Provider or model developer does not permit, or use a VPN or proxy to conceal your location; or

(w) engage in any other activity that exposes Architect, Providers, other users, or payment processors to legal, regulatory, security, or financial harm.

Enforcement. Architect may monitor usage, traffic, payment activity, and technical signals for security, fraud prevention, abuse detection, and compliance. If Architect reasonably believes you have violated this Section 9 or Section 5.4, it may block Requests, suspend or terminate your account, withhold or reverse affected balances, and report the conduct to Providers, payment processors, card networks, law enforcement, or regulators.

10. PROVIDER INFORMATION

Architect displays information about Providers and models, including Quotes, Data Policy Designations, processing locations, and performance and test results. This information is based on Provider disclosures and Architect’s testing, may be delayed, incomplete, or inaccurate, and does not predict future prices or performance. Architect is not responsible for errors in information supplied by Providers.

11. CHANGES TO THE SERVICES; THIRD PARTIES

11.1 We may add, change, limit, or discontinue models, Providers, features, or the Services at any time. No model, Provider, region, price, speed, or capacity is guaranteed. Beta, preview, or experimental features may be incomplete, may change or end without notice, and are used at your own risk.

11.2 The Services rely on third parties, such as Providers and cloud, authentication, and payment providers, and work with third-party tools, such as coding assistants, SDKs, and workflow tools, that are governed by their own terms. Architect is not responsible for third-party services outside its reasonable control and may stop supporting any integration.

11.3 Model names and logos belong to their owners. Displaying them does not imply endorsement by or affiliation with those owners.

12. PRIVACY

Our Privacy Policy describes how we handle personal information. Our Data Processing Agreement, which you accept when you create an account, applies where Architect processes personal information in Customer Content on your behalf, and prevails over these Terms for that processing. You are responsible for choosing models and Routing Rules appropriate for your data. Customer Content and related data may be transmitted to Providers and to the model developers and cloud, security, and payment providers they use, some of which may be located outside your country.

13. CONFIDENTIALITY

Each party will protect the other’s non-public information that is marked or reasonably understood to be confidential, using at least reasonable care, and will use it only to perform under these Terms. This does not apply to information that is public, already known to the receiving party, independently developed, or lawfully received from a third party. A party may disclose confidential information when required by law, after giving notice where legally permitted.

14. OWNERSHIP, FEEDBACK, AND COPYRIGHT

14.1 Architect and its licensors own the Services, software, documentation, and related technology, excluding Customer Content. You receive only the limited right to use the Services under these Terms, and may not copy, modify, reverse engineer, or create derivative works of them, except where the law prohibits this restriction.

14.2 We may use your feedback without restriction or payment to you, and will not identify you publicly as its source without your permission.

15. SUSPENSION AND TERMINATION

15.1 You may stop using the Services at any time, and may close your account by contacting us at [email protected].

15.2 Architect may suspend or terminate your account or access to the Services at any time, for any reason or no reason, with or without notice, and is not liable for doing so.

15.3 If Architect terminates your account for any reason, or if you close your account for any reason, or unused Credits are forfeited, except as required by law.

15.4 On termination, your right to use the Services ends. Accrued payment obligations and Sections 5, 13, 14, and 16 through 20 survive.

16. DISCLAIMERS

TO THE FULLEST EXTENT PERMITTED BY LAW, THE SERVICES, MODELS, PROVIDERS, PROVIDER INFORMATION, AND OUTPUTS ARE PROVIDED “AS IS” AND “AS AVAILABLE.” ARCHITECT DISCLAIMS ALL WARRANTIES, WHETHER EXPRESS, IMPLIED, OR STATUTORY, INCLUDING WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE, NON-INFRINGEMENT, AND ACCURACY, AND ANY WARRANTY THAT THE SERVICES WILL BE UNINTERRUPTED, SECURE, OR ERROR-FREE. ARCHITECT MAKES NO WARRANTY ABOUT ANY PROVIDER’S OR MODEL DEVELOPER’S DATA HANDLING, RETENTION, TRAINING, OR SECURITY PRACTICES, EXCEPT AS EXPRESSLY STATED IN THESE TERMS, OR THAT OUTPUTS ARE SUITABLE FOR ANY REGULATED, HIGH-RISK, OR SAFETY-CRITICAL USE. SOME JURISDICTIONS DO NOT ALLOW THESE EXCLUSIONS, SO SOME MAY NOT APPLY TO YOU.

17. LIMITATION OF LIABILITY

17.1 TO THE FULLEST EXTENT PERMITTED BY LAW, NEITHER PARTY WILL BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, EXEMPLARY, OR PUNITIVE DAMAGES, OR FOR LOST PROFITS, REVENUE, GOODWILL, OR DATA, EVEN IF ADVISED OF THEIR POSSIBILITY.

17.2 EACH PARTY’S TOTAL LIABILITY ARISING OUT OF OR RELATING TO THESE TERMS OR THE SERVICES WILL NOT EXCEED THE GREATER OF (A) THE AMOUNTS YOU PAID TO ARCHITECT FOR THE SERVICES IN THE 12 MONTHS BEFORE THE EVENT GIVING RISE TO THE CLAIM AND (B) US $100.

17.3 THESE LIMITS DO NOT APPLY TO YOUR PAYMENT OBLIGATIONS, INDEMNITY OBLIGATIONS, BREACH OF SECTION 13, FRAUD OR WILLFUL MISCONDUCT, OR LIABILITY THAT CANNOT BE LIMITED BY LAW. THEY ARE AN ESSENTIAL PART OF THE BARGAIN AND APPLY EVEN IF A LIMITED REMEDY FAILS OF ITS ESSENTIAL PURPOSE.

18. INDEMNITY

18.1 You will defend and indemnify Architect and its affiliates, and their officers, directors, employees, and agents, against all claims, losses, liabilities, damages, and expenses, including reasonable legal fees, arising out of third-party claims relating to: (a) your access to or use of the Services, including your Customer Content and use of Outputs; (b) your products or services; (c) your violation of these Terms or applicable law; (d) your violation of any third-party right, including intellectual property, privacy, and publicity rights; (e) any dispute between you and any third party; or (f) misuse of your account, API keys, or payment credentials.

18.2 The indemnified party will notify the indemnifying party promptly and cooperate reasonably. Architect may, at its own expense, assume the exclusive defense of any claim you must indemnify, and you will cooperate with that defense. Neither party may settle a claim in a way that admits fault by, or imposes obligations on, the other without its written consent.

19. GOVERNING LAW AND DISPUTES

19.1 Governing law. These Terms are governed by the laws of the State of New York, without regard to conflict-of-law rules. The Federal Arbitration Act governs this Section 19.

19.2 Arbitration. Except as stated in Section 19.3, you and Architect will resolve any dispute arising out of or relating to these Terms or the Services, whether based in contract, tort, statute, or any other legal theory, and whether it arises during or after the term of these Terms, by binding individual arbitration. The American Arbitration Association (“AAA”) will administer the arbitration under its Commercial Arbitration Rules or, if you are an individual using the Services for personal purposes, its Consumer Arbitration Rules. A single arbitrator will conduct the arbitration in New York County, New York, or by video or on written submissions where the AAA rules permit, and will issue a reasoned written decision. YOU AND ARCHITECT EACH WAIVE THE RIGHT TO A TRIAL BY JURY.

19.3 Exceptions. Either party may (a) bring an individual claim in small claims court; (b) bring a matter to the attention of a government agency; or (c) seek injunctive or other equitable relief in court to protect its intellectual property, confidential information, or the security of the Services.

19.4 Informal resolution. Before starting an arbitration, a party must send the other a written notice describing the claim and the relief sought, and the parties will try in good faith to resolve it for thirty (30) days. Notices to Architect must be sent as described in Section 20.2.

19.5 No class actions. Each party may bring claims only in its individual capacity, and not as a plaintiff or class member in any class, collective, or representative proceeding. The arbitrator may not consolidate the claims of more than one person without both parties’ consent. If this Section 19.5 is found unenforceable for a claim, that claim will proceed in the courts described in Section 19.7, not in arbitration.

19.6 Fees. Arbitration fees are paid as the AAA rules provide.

19.7 Courts. Any claim that is not arbitrated will be brought exclusively in the state or federal courts located in New York County, New York, and each party consents to their jurisdiction. The jury waiver in Section 19.2 also applies in court.

20. GENERAL

20.1 Changes to these Terms. We may update these Terms by posting a new version with a new effective date. For material changes, we will give at least 10 days’ notice by email or in the Services, unless a change is required by law. Changes apply going forward and continuing to use the Services after they take effect means you accept them. Disputes are governed by the version in effect when the dispute arose.

20.2 Notices. We may send notices to your account email or through the Services, and you agree that electronic communications satisfy any legal requirement that they be in writing. Legal notices to Architect must be sent to 1 E Erie St, Suite 525-5211, Chicago, IL 60611 with a copy to [email protected].

20.3 Assignment. Neither party may assign these Terms without the other’s written consent, except to an affiliate or to a successor in a merger, acquisition, or sale of substantially all relevant assets, with notice to the other party.

20.4 Force majeure. Neither party is liable for delays or failures caused by events beyond its reasonable control, such as natural disasters, war, labor disputes, government action, or widespread internet, power, or cloud outages. This does not excuse payment obligations.

20.5 Other terms. The parties are independent contractors. These Terms and the documents they reference are the entire agreement about the Services; purchase-order terms do not apply. If a provision is unenforceable, the rest remains in effect, and not enforcing a provision is not a waiver. These Terms create no third-party beneficiaries.