Terms of service

These Terms of Service (the “Terms”) are a binding agreement between Standard Thinking, Inc. (“Standard Thinking,” “we,” “us,” or “our”) and the person or entity that accesses or uses the Services (“you” or “Customer”).

Zero Data Retention (ZDR) covers Standard API inference, including Standard One and Product Plan inference when activated for you. Coding Plan has separate training and retention terms.

By creating an account, using an API key, signing an order, or otherwise accessing the Services, you agree to these Terms and our Privacy Policy.

1. Agreement and eligibility

1.1 The Services

“Services” means our websites, dashboard, application programming interfaces, software development kits, documentation, model routing and inference services, hosted or dedicated compute, support, and any related product or service we make available.

1.2 Authority and age

You must be at least 18 years old and legally able to enter into this agreement. If you use the Services for an organization, you represent that you have authority to bind that organization, and “you” includes that organization and its authorized users.

1.3 Contract priority

If you and Standard Thinking sign an order form, enterprise agreement, data processing addendum (“DPA”), or other written agreement covering the same Services, that agreement controls to the extent of a conflict. Otherwise, these Terms apply.

Some Services have additional terms (“Service-Specific Terms”) that form part of these Terms when applicable to you. Subject to the customer agreements described above, Service-Specific Terms control over these Terms for the Service they cover. They do not override an applicable ZDR commitment or customer agreement, and any model-training use must satisfy Section 5.3 and the applicable data-processing, restricted-data, and security provisions in Sections 5, 6, and 10.

2. Access to the Services

Access to the Services may require an invitation, selection, capacity, or separate activation. We may determine whether and when to offer access, and the available features, models, configuration, and price, subject to applicable law and any existing agreement. An access inquiry or interest registration does not guarantee access, invitation timing or order, features, or price; it is not a purchase or automatic enrollment. Any paid-use terms and price will be presented before you commit to paid use. Once access is granted, we grant you a limited, non-exclusive, non-transferable, and revocable right to use the Services within the permitted scope of your product and plan.

Coding Plan, Standard API, and Product Plan may be integrated into products you make available to your own users, subject to the selected model’s terms, the applicable Service-Specific Terms, and any applicable order.

You are responsible for your applications, end users, configurations, model selections, and all activity conducted through your account. You must provide any notices and obtain any consents required for your use of the Services.

We may add, modify, deprecate, or discontinue features, models, regions, limits, or interfaces. When commercially reasonable, we will give advance notice of a material breaking change to a generally available paid API. Preview, beta, evaluation, and experimental features may change or end at any time and are provided without production commitments.

3. Accounts and API keys

We use authentication credentials and relevant account data to authenticate requests and control access. You must:

  • Provide accurate account and billing information and keep it current.
  • Protect passwords, API keys, access tokens, and other credentials. Do not publish them or share them with an unrelated third party.
  • Use reasonable controls to limit key access, rotate credentials, and monitor usage.
  • Notify us promptly at contact@standardthinking.ai if you suspect unauthorized access.

You are responsible for charges and activity associated with credentials issued to your account until you notify us of compromise. We may require verification before restoring access or changing account ownership.

4. Inputs, outputs, and ownership

4.1 Customer Content

“Input” means data, text, code, images, audio, video, model files, or other material you submit to the Services. “Output” means material returned by the Services in response to Input. Together, Input and Output are “Customer Content.” Inputs include prompts, messages, code, files, embeddings, media, model artifacts, and retrieval data. Outputs include tokens, text, code, media, classifications, embeddings, and other results generated in response to Inputs.

4.2 Your rights

As between you and Standard Thinking, you retain your rights in Input and, to the extent permitted by applicable law, own Output. You grant us a limited, worldwide license to host, copy, transmit, and process Customer Content only as needed to provide, secure, support, and comply with law in connection with the Services and, only as permitted under Section 5.3, to use retained Customer Content for the model-training purposes specified in the applicable Service-Specific Terms.

4.3 Your responsibilities

You represent that you have all rights, permissions, and lawful bases needed for us to process Input as instructed by you. You are responsible for Customer Content and for deciding whether Output is suitable for your use. Output may not be unique, and other users may receive the same or similar results.

5. Data processing, retention, and model training

5.1 Processing roles, purposes, and service-specific retention

For account, billing, website, security, and business operations, Standard Thinking generally acts as a controller or business. For personal data contained in Customer Content that a business customer submits for inference or customer-enabled storage features, Standard Thinking generally acts as that customer’s processor or service provider and processes the data on documented instructions. Where Service-Specific Terms permit us to use Coding Plan content to train and evaluate our own models, Standard Thinking acts as the controller or business for that use, within the content, purposes, retention terms, and controls stated in the Coding Plan terms.

The customer determines the lawfulness, purpose, and content of its Inputs and is responsible for notices, consents, and instructions to its own users and data subjects.

Processing purposes.

We use authentication credentials and relevant account data to authenticate requests and control access. We process Inputs and Outputs to run the requested model or deployment, route traffic, return results, provide customer-enabled features, protect the platform, and deliver requested support, subject to the processing and retention limits in this Section 5.

We process Customer Content on your behalf to provide, operate, secure, and support the Services you request, and as required by applicable law. We do not use Customer Content for independent purposes unless separately and expressly authorized by you or expressly permitted under Section 5.3, subject to any applicable zero data retention commitment. Coding Plan training and evaluation use, where permitted, is governed by the Coding Plan terms and Section 5.3.

We process account data and service metadata for metering, billing, reliability, capacity planning, incident investigation, abuse prevention, and legal compliance.

We do not use Inputs or Outputs from ZDR-covered inference to train models. De-identified operational statistics and content-free service metadata may be used to improve platform reliability, routing, and performance.

Temporary processing, caching, and ZDR.

Processing may include temporary buffering, queuing, inference caching, and reuse of intermediate results across requests. Processing and retention may vary by service, feature, and configuration, subject to this Section 5 and any applicable customer agreement. ZDR applies only where expressly identified for the relevant service or configuration; these Terms do not override a stricter signed commitment.

For ZDR-covered inference, we do not log Inputs, Outputs, prompts, or content-bearing derivatives. Non-volatile copies of that data are prohibited except for the limited cache offloading described below. Prompts, derived inference caches (including KV caches), and intermediate processing data may remain in volatile memory as reasonably needed for supported inference caching. Access to covered inference content and caches is limited to systems used to deliver the covered service.

Derived KV tensors and hashed lookup keys may be temporarily offloaded to SSD. Raw prompt or output text, including excerpts and token IDs, is not written to SSD. Cache data in memory or on SSD is kept only as long as reasonably needed for supported inference caching and is cleared when no longer needed. It is not retained indefinitely, backed up, or used for model training or other independent purposes. No fixed cache lifetime or deletion immediately after a response is promised.

The temporary caching and limited SSD offloading described above are permitted inference processing under this Section 5.1, subject to any stricter applicable ZDR commitment. Features that retain retrievable content or stored responses outside that scope are subject to separate retention terms. Standard API (including Standard One) inference is covered by ZDR; Product Plan inference is covered when that service is activated for you under this Section 5.1.

A ZDR commitment applies to the identified inference service and configuration. It does not extend to separately enabled storage features, material you voluntarily provide for support or feedback, or copies kept by your application or other services you use. Content-free service metadata is handled under Section 5.2. These Terms do not establish a ZDR commitment for every service or configuration.

5.2 Account data and content-free service metadata

  • Account data: users, organization, permissions, projects, billing contacts, authentication, and support records.
  • Service metadata: model identifier, token counts, timestamps, latency, region, routing decisions, status and error codes, feature use, non-reversible safety signals that cannot reproduce Customer Content, and account or project identifiers. It does not include the substance of Inputs or Outputs. We retain it to meter, bill, operate, protect, and improve the Services.

For Standard API, including Standard One, and Product Plan, we retain content-free service metadata in identifiable form for up to 90 days from the original event for metering, billing, reliability, security, and service operations. Copying, moving, or reprocessing does not restart the period, and identifiable copies are subject to the same limit. After 90 days, we retain such metadata only in sufficiently de-identified form that cannot reasonably identify or be linked to an individual or customer account. We do not attempt to re-identify it or permit recipients to do so. We may retain that de-identified information indefinitely for operational analysis, capacity planning, security, and service improvement. Information we cannot sufficiently de-identify is deleted by the end of the 90-day period, except where a specific legal preservation requirement applies and only for as long as needed for that purpose. Account, billing, tax, and support records are separate categories governed by their own purposes and retention criteria; service event logs are not reclassified into those categories to avoid this limit. This metadata does not include the substance of Inputs or Outputs or content-bearing caches.

5.3 Optional storage, deletion, and model training

The Product Plan data platform is not yet released. Once it is, storage through it is optional. When you enable it, we retain the data you designate to provide that feature on your behalf, outside the scope of inference ZDR. Other customer-enabled features requiring retained content, including history, response caching, batch jobs, retrieval stores, evaluations, fine-tuning, and stored model assets, follow their disclosed retention terms. Content voluntarily provided for support or feedback is also handled separately. Enabling storage does not remove a ZDR commitment from separately covered inference processing.

Retention periods, access and deletion controls, and the effect of disabling a storage feature are described in the product, documentation, or customer agreement. Disabling a feature does not by itself require immediate deletion of previously stored data or its backups. Enabling storage does not by itself authorize training on your content. Any training use for a non-ZDR workflow requires separate, explicit authorization or express permission under Section 5.3. This does not permit training on ZDR-covered inference content.

Unless a specific period is stated, retained Customer Content is kept only as long as reasonably needed for the customer-directed function, requested support, or applicable legal obligations. Retention for Coding Plan model training and evaluation is governed by the applicable Service-Specific Terms, subject to their exclusions and the restrictions in this Section 5.3 on backup use and legal preservation. Any copy preserved under a specific legal requirement is isolated and not used for training or evaluation. When the applicable period ends, data is deleted or returned through the relevant service processes. Deletion may take a reasonable period to propagate through systems, and backup copies may remain until the applicable replacement or deletion cycle, subject to law and contractual commitments. Remaining copies are restricted to recovery, security, or legal preservation purposes and are not used for independent purposes.

Legal preservation requirements may apply to information already in our possession. These provisions do not authorize collecting content records that an applicable ZDR commitment prohibits or overriding a stricter signed agreement.

These provisions remain subject to any stricter signed commitment.

Using a stateful feature does not by itself authorize model training. We do not use Input or Output from inference covered by a ZDR commitment to train models, including where a customer separately opts in. A separate opt-in may apply only to an expressly identified, non-ZDR training workflow under its own terms. Optional storage or a training submission does not remove ZDR protection from separately covered inference.

Where Service-Specific Terms permit that use, they must identify the covered Service and content, the training purposes, the retention period or, where that cannot be specified, the objective criteria used to determine retention, and any controls available to you and their effects. That permission applies only after those terms become effective for you and only to content submitted after that time, subject to any notice or authorization required by applicable law or your agreement. Publishing Service-Specific Terms alone does not authorize that use.

Service-Specific Terms do not authorize training on content covered by a ZDR commitment, restricted data described in Section 6, or content whose use would violate an applicable customer agreement or applicable law.

5.4 Infrastructure, model services, and subprocessors

We may engage cloud infrastructure, networking, model execution, observability, authentication, support, and other service providers to process data on our behalf. Subprocessors must follow the processing and retention restrictions applicable to the relevant service. Where a service or configuration carries a ZDR commitment, subprocessors handling its covered inference content must follow materially equivalent restrictions.

A customer-selected model may be subject to model-specific terms or processing constraints. Where an enterprise agreement provides a subprocessor list or notice mechanism, we will follow that agreement.

Our Privacy Policy explains how we handle personal information.

5.5 Location and transfers

Data may be processed in the United States and other locations where we or our providers operate. If applicable law requires a transfer mechanism, the parties may enter into a DPA incorporating standard contractual clauses or another recognized safeguard. Regional or dedicated processing is available only when expressly included in an order.

5.6 Data subject requests and legal demands

When we act as a processor, we will provide reasonable assistance for a customer’s verified data subject request as required by the applicable DPA and law. We may direct an individual to the customer that controls the data.

We review government and legal demands for validity and scope. Where legally permitted, we notify the affected customer before disclosing Customer Content and seek to limit disclosure to what is required.

6. Acceptable use

Do not submit payment-card data, protected health information, government-classified information, export-controlled technical data, biometric identifiers used for identification, or other specially regulated data unless Standard Thinking has expressly agreed in writing to support that data type and the required safeguards are enabled.

You may not use the Services, directly or through an application, to:

  • violate applicable law, sanctions, export controls, or another person’s privacy, publicity, contractual, or intellectual property rights;
  • generate, distribute, or facilitate malware, credential theft, unauthorized intrusion, denial-of-service activity, spam, fraud, or other harmful or deceptive conduct;
  • exploit or sexualize children, facilitate human trafficking, promote violent wrongdoing, or create content whose distribution is unlawful;
  • circumvent access controls, safety controls, rate limits, quotas, usage restrictions, or billing mechanisms;
  • probe, scrape, reverse engineer, extract, reconstruct, or attempt to discover model weights, source code, training data, non-public prompts, or other proprietary components, except where applicable law expressly permits and cannot be waived;
  • misrepresent Output as human-generated when disclosure is required, impersonate a person without authorization, or use deceptive synthetic media in violation of law;
  • make a fully automated decision that determines a person’s eligibility for employment, housing, credit, insurance, healthcare, legal services, education, or an essential public benefit without legally required review, notice, testing, and appeal safeguards; or
  • submit restricted data described in Section 6 unless a written agreement with us expressly permits that processing.

You must maintain reasonable oversight of your application and its users. We may use automated or manual measures to detect abuse, enforce technical limits, investigate suspected violations, and protect the Services.

7. Models and third-party services

The Services may provide access to models, software, datasets, hosting, payment processing, or other technology licensed or operated by third parties. A model may have its own license, use restriction, or attribution requirement. You are responsible for reviewing and complying with any terms presented for the model you select.

Model availability, behavior, context limits, pricing, and capabilities may change. Standard Thinking does not control third-party model weights or guarantee that a particular model will remain available. We may replace, restrict, or remove a model when required by its licensor, law, security, capacity, or operational considerations.

8. Usage limits and availability

Plans may include request, token, concurrency, throughput, region, or other limits. You may not evade limits by creating additional accounts or coordinating requests across accounts. We may throttle or reject traffic that exceeds documented limits or threatens the reliability or security of the Services.

Coding Plan access configuration and any service credits are set out in the Coding Plan terms in Section 20 and any applicable invitation or signed agreement.

A Product Plan order specifies the guaranteed base allocation for each model, qualifying workload, measurement conditions, and applicable remedies. Free extra capacity above that base may be provided while available and is not reserved. Auto scaling requires your authorization and available capacity. An availability SLA does not by itself create a latency or throughput commitment.

Model availability depends on capacity and can change at any time. A listed model may be unavailable or not yet released. Plans and orders state their own commitments.

Unless your order or the applicable Service-Specific Terms expressly include a service level commitment, the Services are provided on a commercially reasonable, best-effort basis without guaranteed uptime, latency, or throughput. Our Service Level Agreement in Sections 8.1 through 8.7 applies only where an eligible order incorporates it.

8.1 Service Level Agreement

This Service Level Agreement (“SLA”), consisting of Sections 8.1 through 8.7, applies only when a signed Standard Thinking order form expressly incorporates it and identifies a covered service, monthly availability commitment, and credit schedule.

Public and evaluation plans. Unless an order or applicable Service-Specific Terms state otherwise, pay-per-token (Standard API), free, trial, preview, beta, and evaluation services are provided on a best-effort basis and do not carry a service credit. Coding Plan has no public SLA or service-credit program unless a signed agreement expressly provides one.

8.2 Definitions

  • Covered Service means the production API gateway or Product Plan base allocation identified in the applicable order.
  • Monthly Uptime Percentage means total minutes in the calendar month minus Unavailable Minutes, divided by total minutes in that month, multiplied by 100.
  • Unavailable Minute means a full minute during which the Covered Service cannot successfully accept and process valid requests, as measured by Standard Thinking.
  • Service Credit means a credit calculated under the schedule in the applicable order and applied to a future invoice for the affected Covered Service.

8.3 Measurement

Availability is measured from Standard Thinking’s monitoring at the service boundary for the covered region and endpoint. A partial degradation, elevated latency, individual model error, quota rejection, malformed request, or failure outside that boundary is not downtime unless the order expressly says otherwise.

Availability and throughput are measured separately. A Product Plan order defines its base-throughput commitment, measurement conditions, and remedies. This Section 8 alone does not establish those commitments, and free extra capacity is not guaranteed.

Where a Covered Service spans multiple regions or deployments, availability may be calculated separately for each affected deployment or as specified in the order.

8.4 Exclusions

Unavailable Minutes do not include unavailability caused by:

  • scheduled maintenance announced in advance when reasonably practicable, or emergency maintenance needed to protect the Services;
  • Customer Content, code, configuration, credentials, networks, dependencies, rate-limit violations, or failure to follow documentation;
  • third-party models, model licensors, internet carriers, cloud or utility failures outside our reasonable control;
  • beta, preview, evaluation, free, or unsupported features;
  • a suspension or restriction permitted by the agreement;
  • force majeure, denial-of-service attacks, widespread internet disruption, government action, or other events beyond our reasonable control; or
  • Customer’s failure to use available redundancy, failover, or supported versions required by the order.

8.5 Service credit requests

To request a Service Credit, the customer must email contact@standardthinking.ai within 30 days after the end of the affected month and include the organization, project, covered endpoint, incident date and time, request identifiers where available, and a description of the impact.

We will evaluate the request against our service records. Approved credits are applied to future fees for the same Covered Service, are not cash refunds, and cannot exceed the affected service fees for the month unless the order provides otherwise.

8.6 Sole remedy

Service Credits are the customer’s sole and exclusive remedy for a failure to meet an availability commitment, except where the applicable order expressly provides another remedy or applicable law does not permit this limitation.

8.7 Order priority

If this SLA conflicts with a signed order or enterprise agreement, the signed agreement controls. Capitalized terms not defined here have the meanings in these Terms or the applicable order.

9. Fees, billing, and taxes

9.1 Fees and payment arrangements

You will pay the fees shown at purchase, in the dashboard, or in an order form. Standard API usage is paid for through prepaid credits unless an order form expressly provides another payment arrangement. Coding Plan subscriptions, Product Plan commitments, and other separately purchased Services follow the payment terms disclosed at checkout or in the applicable order. Unless stated otherwise, amounts are in U.S. dollars and payment obligations are non-cancelable once incurred.

9.2 Prepaid Credits and usage deductions

You purchase service credits in advance, which are added to your account balance (“Credits”). We calculate usage charges from our metering records at the rates applicable when the usage occurs and deduct those charges from your Credits. Buying Credits does not fix the price of future usage. Credits may be used only for Services designated as payable with Credits and cannot be transferred between accounts or redeemed for cash except under Section 9.4.

Prepaid balance is not a hard spending cap: metering and charges for accepted requests may be finalized after the request completes or terminates and may result in a negative balance. Any usage already incurred remains payable. We may reject new requests or suspend prepaid usage when the available balance is insufficient. The displayed balance does not guarantee a request will fail or be cancelled at a particular threshold. Future prices may change prospectively under Section 9.6; purchasing Credits does not lock future rates.

9.3 Optional automatic recharge

If we offer automatic recharge and you choose to enable it, you authorize us and our payment processor to charge your selected payment method for additional Credits according to the recharge amount and balance threshold you select. Automatic recharge requires your separate activation. You may change or disable future automatic recharges through the available billing controls or by contacting us. Disabling automatic recharge does not reverse a charge already initiated or cancel usage charges already incurred.

9.4 Refunds, expiration, and account closure

Payments, including purchases of Credits and unused Credit balances, are non-refundable except as required by applicable law or expressly stated in the terms disclosed at purchase or in your order. Any expiration date for purchased Credits must be disclosed before purchase. If no expiration date is disclosed, those Credits do not expire. We will not retroactively shorten the disclosed validity period of Credits you have already purchased. If we permanently discontinue a paid Service for reasons other than your breach, any applicable rights concerning remaining purchased Credits for that Service will be addressed under the terms disclosed at purchase, your order, and applicable law. Promotional Credits remain subject to their disclosed terms.

Closing your account or terminating the Services does not automatically entitle you to a refund of unused Credits or cancel charges already incurred. Any remaining Credit balance is subject to the refund and expiration terms above, and account closure does not override a refund right required by applicable law or expressly provided in your purchase terms or order.

9.5 Payment authorization, taxes, and disputes

You authorize us and our payment processor to charge your payment method for Credit purchases you initiate, automatic recharges you enable, applicable taxes, and other fees or recurring subscriptions disclosed and authorized at checkout or in your order. You are responsible for taxes other than taxes on our net income. You must raise a good-faith billing dispute within 30 days after the applicable invoice, payment charge, or Credit deduction.

9.6 Pricing changes and overdue invoices

We may change public pricing prospectively by posting or providing notice. A price fixed for a committed order term will not change during that term unless the order says otherwise. For separately agreed invoiced payments, overdue undisputed amounts may accrue interest at the lesser of 1.5% per month or the maximum lawful rate, and we may suspend paid features after reasonable notice.

10. Confidentiality and security

Each party may receive non-public information that a reasonable person would understand to be confidential. The receiving party will use that information only to perform or exercise rights under the agreement and will protect it with at least reasonable care. These duties do not cover information that is public without breach, already lawfully known, independently developed, or rightfully received without restriction.

A party may disclose confidential information when legally required after giving notice where permitted and reasonable assistance at the disclosing party’s expense. Customer Content is your confidential information, subject to the processing rights in these Terms.

10.1 Security safeguards

We use technical and organizational measures designed to protect data against unauthorized or unlawful access, use, alteration, and loss. Measures are selected based on the service, sensitivity, and risk and may include encryption, access controls, content-free operational logging, monitoring, environment separation, and incident response. Personnel and providers with access to Customer Content or other data we process are subject to confidentiality obligations.

Sections 10.2 through 10.8 give the public overview. Specific enterprise controls belong in the applicable security exhibit or order.

Standard Thinking uses layered safeguards designed to protect the confidentiality, integrity, and availability of customer workloads. Security is a shared responsibility between our platform and each customer deployment.

10.2 Data protection

  • Network traffic to supported endpoints is encrypted in transit using current transport security protocols.
  • Stored customer and account data is encrypted at rest where appropriate to the storage system and risk.
  • Customer Content is processed under the retention and caching terms applicable to the service and configuration. ZDR commitments apply where expressly identified; customer-enabled storage follows its separate retention terms. See Section 5.
  • API and Product Plan content-free service metadata is identifiable for up to 90 days from its original event and is then retained only in sufficiently de-identified form under Section 5.2. Account, billing, tax, and support records follow separate criteria. Customer Content remains subject to the purpose and retention limits in these Terms.
  • Content from Standard API and Product Plan inference covered by ZDR is not used to train models, including where a customer separately opts in. Coding Plan training and evaluation use is governed by its Service-Specific Terms.
  • Subprocessors are required to follow the processing and retention restrictions applicable to the service, including materially equivalent ZDR restrictions where a ZDR commitment applies.

10.3 Identity and access

We design access to production systems around least privilege, separation of duties, and authenticated administrative access. Access to sensitive systems is limited to personnel and services with a business need and is reviewed as roles change. Administrative and customer account activity may be logged for security and audit purposes.

Project-scoped credentials and revocation controls help customers limit the effect of a compromised key. Enterprise configurations may add organization controls, single sign-on, private networking, dedicated capacity, or other safeguards described in an order.

10.4 Operational security

Our operational program is designed to include secure configuration, change control, dependency and vulnerability management, environment separation, backup and recovery planning, and monitoring for suspicious behavior. We evaluate reported vulnerabilities and prioritize remediation based on exploitability and potential impact.

We use infrastructure and model partners to provide portions of the Services. We assess vendors based on the nature of the service and data involved and use contractual and technical controls appropriate to the risk.

10.5 Incident response

We maintain procedures designed to identify, contain, investigate, remediate, and learn from security incidents. If an incident affects customer data, we will notify affected customers as required by applicable law and contractual commitments, and we will provide information reasonably available to support their response.

Service availability events are communicated through our status page when appropriate.

10.6 Customer responsibilities

Customers must:

  • protect account credentials and API keys, limit access, rotate secrets, and avoid embedding secrets in public client code;
  • configure applications, networks, permissions, and data retention according to their risk and legal obligations;
  • validate Output before using it in production or a consequential decision;
  • avoid sending restricted data described in Section 6 unless a written agreement permits it;
  • monitor for unauthorized activity and promptly disable compromised credentials; and
  • keep integrations and dependencies current and follow our documentation and usage limits.

10.7 Reporting a vulnerability

If you believe you found a security vulnerability, email contact@standardthinking.ai with a concise description, affected endpoint or feature, reproduction steps, and potential impact. Do not access data that is not yours, disrupt production, use social engineering, or publicly disclose an unresolved issue.

We will acknowledge legitimate reports and coordinate remediation and disclosure as appropriate. This section does not create a bug bounty or promise payment.

10.8 Scope and updates

This section is a public overview and does not replace a signed security exhibit, data processing addendum, or enterprise order. Safeguards evolve with the Services and threat landscape, so we may update this section as our program changes.

11. Standard Thinking property and feedback

We and our licensors retain our respective rights in the Services, including our software, APIs, interfaces, documentation, routing systems, models we develop and own, usage analytics, trademarks, and related intellectual property. These Terms grant only the access rights expressly stated above.

This section does not change your rights in Customer Content under Section 4. Ownership and use rights in customer-specific fine-tuned models, adapters, and other model deliverables are governed by the applicable written agreement and any underlying model license. Developing or fine-tuning a model for a customer does not, by itself, give Standard Thinking ownership of all rights in that model or its deliverables.

If you provide feedback, you grant Standard Thinking a perpetual, irrevocable, worldwide, royalty-free right to use and incorporate it without restriction or compensation. Do not provide feedback that you are not authorized to share. Customer Content stays subject to Sections 4 and 5, even when you attach it to feedback or a support request.

12. Suspension and termination

You may stop using the Services at any time and may close your account subject to outstanding payment obligations and any committed order term. We may suspend or limit access if we reasonably believe that your use violates these Terms, creates a security or legal risk, harms another customer, threatens service integrity, or involves overdue undisputed fees.

When practicable, we will give notice and an opportunity to cure. We may act immediately when necessary to prevent harm or comply with law. Upon termination, your access rights end. Provisions that by their nature should survive—including payment, ownership, confidentiality, disclaimers, liability limits, indemnity, and dispute terms—will survive.

13. AI-specific limitations

Machine-learning systems are probabilistic. Output may be inaccurate, incomplete, biased, offensive, out of date, insecure, or unsuitable for your purpose. It may contain fabricated facts or code defects and may resemble content generated for another user.

You are responsible for testing and evaluating models, applying appropriate human review, and implementing safeguards proportionate to your use case. Do not rely on Output as a sole source of truth or as professional medical, legal, financial, safety, or other expert advice. We do not warrant that Output is accurate, unique, non-infringing, or fit for a particular use.

14. Disclaimers

We do not warrant that the Services will be uninterrupted, error-free, secure, or free of harmful components, or that defects will be corrected. Any warranty or remedy expressly stated in a signed order is limited to that order. These disclaimers do not apply where prohibited by law.

15. Limitation of liability

The exclusions and cap apply regardless of the theory of liability and even if a remedy fails of its essential purpose. They do not limit liability that cannot be limited by law. A signed enterprise agreement may provide different limits.

16. Indemnification

You will defend, indemnify, and hold harmless Standard Thinking, its affiliates, and their personnel from third-party claims, damages, losses, and reasonable costs arising from your Customer Content, your applications or end users, your violation of these Terms or law, or your infringement or misappropriation of another party’s rights.

We will provide prompt notice and reasonable cooperation, and you may control the defense and settlement, provided that you may not admit fault or impose an obligation on us without our written consent. Any Standard Thinking indemnity will apply only if included in a signed order or enterprise agreement.

17. U.S. trade controls

You will comply with applicable U.S. and other export-control, import, sanctions, and anti-boycott laws. You represent that neither you nor your authorized users are prohibited parties or located in a jurisdiction where use of the Services is barred. You may not use the Services for a prohibited end use or provide access in violation of restrictions administered by the U.S. Department of Commerce, U.S. Department of the Treasury, or another applicable authority.

18. Governing law and disputes

Before filing a claim, each party will make a good-faith effort for 30 days to resolve the dispute through written notice. These Terms are governed by the laws of the State of Delaware, without regard to conflict-of-law rules. The United Nations Convention on Contracts for the International Sale of Goods does not apply.

Except where applicable law requires otherwise, the state and federal courts located in Delaware will have exclusive jurisdiction over disputes arising from these Terms or the Services, and each party consents to personal jurisdiction and venue in those courts.

19. Changes and general terms

The October 13, 2026 revision incorporates the former Data policy into these Terms and updates the temporary inference caching and access provisions in Section 5. This revision does not restart or extend any applicable retention period or change the terms applicable to earlier content or billing periods.

We may update these Terms by posting a revised version and changing the “Last Updated” date. If a change materially reduces your rights, we will provide reasonable advance notice through the Services or the email associated with your account. Continued use after the effective date constitutes acceptance.

Neither party is liable for delay or failure caused by events beyond its reasonable control. You may not assign these Terms without our consent, except in connection with a merger or sale of substantially all relevant assets after notice to us. We may assign these Terms in connection with a reorganization, merger, acquisition, or sale of assets.

If a provision is unenforceable, it will be modified only as needed and the remainder will continue in effect. A failure to enforce a provision is not a waiver. These Terms and incorporated agreements are the complete agreement regarding the Services and supersede prior discussions on that subject.

Contact

Questions, legal notices, DPA requests, and inquiries about regional processing or retention controls may be sent to contact@standardthinking.ai.

20. Coding Plan terms

This Section 20 sets out the Coding Plan terms, which are Service-Specific Terms under Section 1.3. This revision is posted October 5, 2026. Changes to model-training retention apply only to content received after these revised terms become effective for you in accordance with the Terms and any required notice or authorization. Content received earlier, and rights or claims arising under terms applicable to an earlier billing period, remain governed by those earlier terms. Where this Section 20 is silent, the other sections of the Terms control.

20.1 Scope

Coding Plan is a subscription for one developer. The models currently listed for the plan are Kimi K3, GLM-5.3, and DeepSeek V4.1 Flash. We may add, replace, update, restrict, or remove models and change their versions, features, or limits for licensing, legal, security, capacity, or operational reasons. A listing describes the current lineup and does not commit us to provide a particular model or fixed number of models throughout a subscription term. Changes remain subject to applicable law, your agreement, and notice requirements in the Terms. Section 7 applies to these models.

Each subscription is for one person. Do not share its credentials with other people. Serving your application’s end users through your own subscription is not sharing credentials, and those requests count toward your capacity. Section 3 applies to your API keys and other credentials.

Coding Plan may serve an application’s end users under the same conditions as Standard API. Those conditions are the selected model’s terms, the acceptable use rules in Section 6, and the capacity described in Access configuration. As Section 2 provides, you are responsible for your applications and end users, including any notices and consents they require.

20.2 Plans and billing

Coding Plan is sold as two plans, Weekly and Monthly. Weekly is $149 per developer for a 7-day billing period. Monthly is $549 per developer for a 30-day billing period. Prices may change prospectively as described in Section 9.6. Both plans currently offer the same model selection and access configuration, subject to these terms and capacity.

Each plan renews automatically at the end of its billing period until you cancel. Payment, taxes, and refunds follow Section 9, including the authorization to charge each renewal in Section 9.5.

You can switch from Weekly to Monthly at any time. We credit the unused days of your current week at your applicable Weekly fee divided by 7 per day against your applicable Monthly fee. At the listed prices, this is $149 divided by 7 per day against $549. Monthly starts on the day of the switch and runs for 30 days. You can switch from Monthly to Weekly only at renewal.

You can cancel at any time. Access continues to the end of the billing period you have paid for, and the plan does not renew. No refund applies beyond Section 9.4.

20.3 Access configuration

There is no monthly token allowance. Concurrency and throughput follow the access configuration for your subscription and may be reduced during peak demand. This configuration does not guarantee reserved slots, availability, throughput, or service at peak times. Your invitation and applicable order confirm the access configuration that applies to you.

Any queueing, concurrency, and other access conditions are specified in the invitation or applicable order. We make no uptime, latency, or throughput promise unless expressly stated in that agreement.

20.4 Service credits

Coding Plan has no contractual service-credit program unless your signed agreement expressly provides one. This change applies only to billing periods that begin after it becomes effective for you in accordance with the Terms. Terms in effect at the start of an earlier billing period continue to govern that entire period, including qualifying requests made later in the period and claims submitted afterward within the applicable claim window. This change does not affect rights or remedies that cannot be excluded under applicable law.

20.5 Model training

Content and purpose. Requests sent through Coding Plan may be used to train and evaluate models developed by Standard Thinking. This covers the Input of those requests and the Output returned for them. You permit this use as this section describes, only after these terms become effective for you and only for content submitted after that time, subject to any notice or authorization required by applicable law or your agreement. This section states the covered content, purposes, retention criteria, and controls required by Section 5.3. We have not started this use. We will note the start date on this page before we begin.

Control. Coding Plan has no setting that turns this use off. Do not submit content you are not authorized to permit for these purposes. Standard API and Product Plan inference content covered by ZDR is not used for training; an access inquiry or use of optional storage is not permission to train on inference content.

Retention. We retain Inputs, Outputs, and content-bearing copies or extracts only while they remain reasonably needed for the permitted training and evaluation activities, including dataset preparation, training, validation, evaluation, and reproducibility. In determining continued need, we consider the content’s nature and sensitivity, its relevance to those activities, preparation or evaluation requirements, and applicable legal and contractual requirements. These terms specify no fixed retention period. When retention is no longer justified for the permitted purposes, we delete the content or, where appropriate and lawful, render it anonymous. Deleting source content does not by itself require retraining or withdrawing a completed model, subject to applicable law.

Exclusions. We do not use content for training or evaluation if it is:

  • covered by a Zero Data Retention (“ZDR”) commitment;
  • restricted data described in Section 6; or
  • content whose use would violate an applicable customer agreement or applicable law.

Providers and your responsibilities. We use this content only to train and evaluate models developed by Standard Thinking. Providers may help perform that work for us under confidentiality, purpose, access, retention, and deletion restrictions. We do not authorize them to use your content to train or improve their own models.

You must have the rights and authority needed to permit this use of the content you send, including content from your own end users. You must give any notice and obtain any consent that applicable law requires. Your permission does not transfer ownership of your content to us, authorize use of another person’s confidential material without that person’s permission, or relieve us of our own obligations under applicable law. If you lack authority to permit training on work-related code or other material, do not send it through Coding Plan.

Scope and changes. This section applies only to Input and Output submitted after these terms become effective for you. It does not apply to earlier requests.

We give notice of material changes as Section 19 provides. Any change applies only after it becomes effective for you in accordance with the Terms and applicable law. Content received earlier remains subject to the retention and use terms that applied when received.

20.6 Order of precedence

If these terms conflict with a signed order, enterprise agreement, or other written agreement that covers Coding Plan, that agreement controls. Otherwise this Section 20 controls over the other sections of the Terms for Coding Plan, within the limits in Section 1.3. Capitalized terms not defined here have the meanings given elsewhere in the Terms.

Contact us about these terms