Terms of service

These terms govern the Experiential Labs platform: the dashboard at platform.experientiallabs.ai, the API at api.experientiallabs.ai, and the API keys, credits, and features you get through them (the “Platform”). The Platform is operated by Resolute Labs AI, Inc., a Delaware corporation doing business as Experiential Labs (“Experiential Labs”, “we”, “us”). The marketing website at experientiallabs.ai has its own terms and privacy policy; those documents stop where the Platform begins, and this one takes over. By creating an account, minting an API key, or sending a request through the gateway, you agree to these terms and to the Platform privacy policy. If you are using the Platform for an organization, you represent that you have authority to bind it, and “you” means that organization.

1. Eligibility and accounts

You must be at least 18 years old. An account belongs to a person and is a member of one or more organizations; an organization is the unit that holds API keys, credits, model access, settings, and billing. The first member of an organization is its admin. Admins can invite members, mint and revoke keys, change settings, and buy credits; every invitation must be accepted by the invitee, and membership is unlimited on every plan. You are responsible for keeping your sign-in methods secure and for everything done through your account and your organization’s keys.

2. API keys and verification

An API key (prefix xpl_) authenticates as your organization. Keep keys secret, put them in server-side configuration, and revoke any key that may have leaked; a revoked key cannot be re-enabled. A key marked as a provisioning key can also create, list, update, and revoke keys through the API; treat it as a server-side secret and never use it for inference traffic. We store only a hash of each key and can never show you a key again after it is created.

New organizations can sign in and list models immediately, but spending credits is locked until the organization is verified. Verification means adding a card and completing a small settled charge (one dollar at the time of writing), or, where we have enabled it, proving control of the sign-up email. Some promotional models also require a saved card and a settled charge. We may ask for additional verification when an account or its traffic looks like abuse.

3. The service

The Platform is a hosted gateway: you call OpenAI-compatible and Anthropic-compatible endpoints with one key, and we route each request to a model provider, stream the response back, and record what it cost. The models available, their prices, and the providers behind them are listed in the catalog and change over time as providers launch, reprice, or retire models; we will not keep a model available that its provider has withdrawn. We may add, change, or remove features. Features marked as beta or preview may change or be withdrawn without notice and are excluded from any commitment we make about the rest of the service.

Unless you have a signed agreement with us that says otherwise, we provide the Platform with commercially reasonable efforts and without a service-level guarantee. The security and reliability page states what the running system does today.

4. Your content

“Customer Content” is what you send through the Platform and what comes back: prompts, messages, tool definitions and calls, files, uploaded traces, model responses, and anything derived from them that is specific to you. You own Customer Content, including model outputs as between you and us, and you are responsible for having the rights and consents needed to send it. You grant us a non-exclusive, worldwide, royalty-free license to host, transmit, store, and process Customer Content as needed to provide the Platform to you, to keep it secure, to enforce these terms, and as described in the privacy policy.

Unless your organization is in private mode, you also grant us a non-exclusive, worldwide, royalty-free license to use the Customer Content the Platform captures for our own purposes, including to develop, evaluate, and improve our services and the models we host. Private mode, available on the Pro plan, ends that license for all of the organization’s content, including content captured before it was turned on, which turning private mode on makes unavailable at once and then erases; nothing new is stored for an organization in private mode. Content of requests served on your own provider keys is never stored and never used this way. We do not sell Customer Content and never show one customer’s content to another. We may compile aggregated statistics about Platform usage (request counts, latency, spend by model) that do not identify you or any person.

5. Content capture and private mode

By default the Platform captures the prompt and the response of the calls your organization makes, keeps them as the privacy policy’s retention schedule describes, uses them to give you per-request logs, prompt grouping in Insights, and the intelligence features, and may use them for our own purposes as described in section 4. This is on for every new organization. Organizations on the Pro plan (and organizations that were verified before this default was introduced) can switch to private mode in Settings, after which no prompt, response, or continuation content is stored by the Platform and none of the organization’s content is used for our own purposes; request metadata (model, tokens, latency, cost, status, key, and the end-user label you send) is still recorded for billing and usage reporting. Private mode is organization-wide, not per key. Requests served on your own provider keys store no content on any plan. The privacy policy states exactly what is stored, for how long, and who can access it.

6. Model providers and bring-your-own-key

Serving your request means sending its content to the model provider that serves it (for example OpenAI, Anthropic, Google, Microsoft Azure, Amazon Bedrock, or a model hosted on our own GPU infrastructure) and returning what it produces. Each provider’s usage policies apply to the requests it serves, and a provider may refuse a request under those policies. You can restrict which providers your organization’s traffic may reach, and require that a provider offer zero data retention or a no-training commitment for your requests, through the provider data controls; those controls fail closed, so a model no compliant provider can serve is refused rather than routed elsewhere.

When you connect your own provider credentials (bring-your-own-key), requests routed on them are billed to you by that provider under your agreement with it. The request still passes through the Platform to reach the provider, but no prompt or response content is stored for it and none is used for our own purposes; only metadata is recorded. We store your provider secrets encrypted and never return them to a browser. You are responsible for the provider account, its terms, and its spend.

7. Credits, plans, and payment

Platform-funded usage is prepaid with credits. One credit is one US cent. Platform-funded model tokens settle at the catalog rates of the serving lane; applicable platform-run search fees and explicitly requested Intelligence model usage also draw credits. BYOK provider usage is billed directly by the provider, and promo-funded usage is not a wallet debit. Purchased top-up credits do not expire and are not subject to subscription rollover limits. Purchases are non-refundable except where the law requires. Free grants and bonuses (including welcome, promotional, and goodwill grants) may carry their own expiry and conditions, are limited to one welcome grant per person, and may be reduced or revoked if obtained through misuse. Payments are processed by Stripe; we never see your full card number.

If you enable auto-recharge, we charge your saved card off-session when your balance falls below the threshold you set; you can change or turn it off at any time in the credits page. Pro is available as a monthly or annual subscription and unlocks features such as private mode and intelligence. Monthly plans grant their allowance on paid billing cycles; annual plans grant the full annual allowance upfront. Subscriptions renew until cancelled. Cancellation takes effect at the end of the paid period and stops future grants; cancellation alone does not remove the remaining balance.

The following subscription rollover policy is prospective and applies only after activation, to eligible grants from the effective date disclosed in your plan or checkout. Publication of these terms does not activate it. Before activation, paid renewals add the full allowance without a rollover cap. Subscription credits granted before activation remain permanently grandfathered outside the capped pool, including paid pre-activation invoices delivered late. Purchased top-ups, free grants, bonuses, and model promotional allowances are excluded from these limits; your total wallet may therefore exceed the subscription ceiling.

At a paid monthly renewal, the eligible subscription balance after the new grant is the lesser of unused eligible credits plus the new allowance or twice the new monthly allowance. At a paid annual renewal into the same or a higher annual allowance, unused eligible credits carry up to 15% of the previous annual allowance, then the full new annual allowance is added. That percentage is not applied to the unused balance. An annual renewal into a lower annual allowance carries no policy-governed annual credits. These are paid-renewal limits, not calendar month-end deductions or a continuously enforced wallet ceiling; failed payments do not trigger them.

A scheduled monthly downgrade uses the new allowance at the next paid renewal, without an immediate deduction. A confirmed full-price upgrade preserves the existing balance and adds the new paid allotment; rollover is evaluated at the following natural renewal, not retroactively at upgrade. At a scheduled monthly-to-annual boundary, up to the outgoing monthly allowance carries before the full annual grant is added. At an annual-to-monthly boundary, no governed annual credits carry. An unexpired annual allotment is not subject to a monthly ceiling. Re-subscribing does not make policy-governed credits grandfathered.

A legitimately paid older invoice delivered late is credited once and does not apply its older ceiling to newer grants; its eligible remainder is considered at the next renewal. A payment restoration keeps its original policy status and follows the same timing rule. Refund and dispute adjustments account for amounts already removed by rollover, without double-removing those credits; this does not prevent recovery of already-spent refunded credits. See the billing guide for examples and the credit consumption order.

Prices, including per-model prices, can change; a change applies to usage after it takes effect. You are responsible for any taxes other than our income taxes. Raise a billing dispute with us within thirty days of the charge; a chargeback opened instead of a dispute may lead to suspension while we resolve it.

8. Free allowances and promotions

Free promotional model access, welcome grants, and rate allowances are offered at our discretion, may carry per-model, per-day, or per-hour limits, and may be changed or ended at any time. They are for one organization per person or business. Creating multiple accounts, rotating email addresses, reusing a card that already verified another organization, or otherwise routing around a limit forfeits the allowance and may lead to suspension.

9. Acceptable use

You agree not to:

• violate any law, including export controls and sanctions, or infringe others’ rights;
• send content or build applications that a serving provider’s usage policy prohibits, or use outputs to violate those policies;
• use the Platform for automated decisions with legal or similarly significant effects on people without human review, or for uses regulated as high-risk without the safeguards the law requires;
• send protected health information, payment card data, or other specially regulated data unless we have agreed to it in writing;
• probe, scan, or test the security of the Platform or its providers without our written authorization, or attempt to access another organization’s data or keys;
• overload the service, evade rate limits or spending controls, or test stolen or unauthorized payment cards;
• misrepresent model output as human-authored where that matters to the audience, or use it to deceive, harass, or harm.

You may build products on the Platform and resell access to your own customers, including by attributing requests to them and rebilling settled cost. You remain responsible to us for your customers’ traffic.

10. Enforcement

We protect the shared provider accounts and the other customers on the Platform. We may rate-limit keys, hold a request, or place an organization under review (“quarantine”) when its traffic draws a high rate of provider policy refusals, when payment activity matches card-testing patterns, when a sign-up email bounces, or when an operator has reasonable grounds. Under review, model calls return an explicit error naming the reason category, purchases may be blocked, and the organization can appeal from the dashboard; we answer appeals as quickly as we can. A paid plan does not exempt an organization from refusal-based review. We may ban accounts that abuse the service, which revokes their keys and ends their sessions; revoked keys stay revoked. We may also refuse sign-ups from disposable or blocked email domains.

11. Intellectual property

We own the Platform and everything in it other than Customer Content: the software, catalog, documentation, routing, pricing, and design, and all improvements to them. The gateway engine is published under the Apache License 2.0 at github.com/experientiallabs/experiential, and that license governs your use of that code. If you give us feedback, we may use it without obligation to you.

12. Confidentiality

Each of us will protect the other’s non-public information with at least reasonable care and use it only for this relationship. This does not cover information that is public through no breach, already lawfully known, independently developed, or lawfully received from someone else. Either of us may disclose when legally required, with notice to the other where the law allows.

13. Your data: export, deletion, and what remains

You can export your usage history at any time through the usage API and the dashboard. An organization admin can delete the organization’s stored content (captured prompts, uploaded traces, batch files, and their derived records) from Settings; that deletion is immediate and cannot be undone. Any member can delete their own account. When an account deletion empties an organization, its API keys are revoked and the organization is removed where that is safe; an organization holding an active subscription is kept until the subscription is cancelled, so that billing can be settled. We keep billing and usage records (amounts, tokens, model, timestamps) after deletion with identifying fields removed, because they are our financial records, and we keep the audit log. The privacy policy has the full schedule.

14. Suspension and termination

You may stop using the Platform at any time. Purchased top-up credits remain available until used; subscription credits follow section 7, including its grandfathering and prospective paid-renewal limits. We may suspend or terminate access for non-payment, for a security risk, for a breach of these terms, or where the law requires, with notice where practicable. On termination we will delete Customer Content in line with the privacy policy’s retention schedule. Sections on payment, content, ownership, confidentiality, disclaimers, liability, and disputes survive termination.

15. Disclaimers

THE PLATFORM IS PROVIDED “AS IS” AND “AS AVAILABLE”. TO THE FULLEST EXTENT PERMITTED BY LAW WE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, NON-INFRINGEMENT, AND ANY WARRANTY ABOUT THE ACCURACY, COMPLETENESS, OR RELIABILITY OF MODEL OUTPUT. Model output can be wrong, biased, or unsafe; you are responsible for evaluating it before relying on it or showing it to your users.

16. Limitation of liability

TO THE FULLEST EXTENT PERMITTED BY LAW, NEITHER PARTY IS LIABLE FOR INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, EXEMPLARY, OR PUNITIVE DAMAGES, OR FOR LOST PROFITS, REVENUE, DATA, OR GOODWILL, HOWEVER CAUSED. EXCEPT FOR YOUR PAYMENT OBLIGATIONS, EITHER PARTY’S BREACH OF CONFIDENTIALITY, INFRINGEMENT OR MISAPPROPRIATION OF THE OTHER’S INTELLECTUAL PROPERTY, YOUR INDEMNITY OBLIGATIONS, FRAUD, OR WILLFUL MISCONDUCT, EACH PARTY’S TOTAL LIABILITY UNDER THESE TERMS IS LIMITED TO THE GREATER OF THE AMOUNTS YOU PAID US IN THE TWELVE MONTHS BEFORE THE EVENT GIVING RISE TO THE CLAIM AND ONE HUNDRED US DOLLARS.

17. Indemnity

You will defend and indemnify Experiential Labs and its officers, directors, employees, and agents against third-party claims arising from Customer Content, your products, or your breach of these terms or of law, including a provider’s usage policy. We will defend and indemnify you against third-party claims that the Platform, as provided by us and used as permitted, infringes that party’s intellectual property, excluding claims arising from Customer Content, from the models and their output, or from combinations with things we did not supply. The indemnified party must give prompt notice, reasonable cooperation, and control of the defense to the indemnifying party.

18. Governing law and disputes

These terms are governed by the laws of the State of Delaware, without regard to its conflict-of-laws rules. Disputes will be brought exclusively in the state or federal courts located in Delaware, and each party consents to their jurisdiction, except that either party may seek injunctive relief anywhere to protect its intellectual property or confidential information.

19. Enterprise agreements

If your organization has a signed order form, master agreement, or data processing agreement with us, that document governs where it conflicts with these terms. Organizations that need a data processing agreement, a security review, custom retention, or contractual commitments beyond these terms can write to legal@experientiallabs.ai.

20. Changes

We may update these terms. For a material change we will notify organization admins by email or in the dashboard at least fourteen days before it takes effect, except where a change is needed sooner for legal or security reasons. Continued use after the effective date is acceptance. If you do not agree, stop using the Platform and, if you have a paid plan, cancel it before the change takes effect.

21. General

These terms and the privacy policy are the entire agreement between you and us about the Platform, unless you have a signed agreement with us. If a provision is unenforceable, the rest stays in force. Neither party may assign these terms without the other’s consent, except to a successor in a merger, acquisition, or sale of substantially all assets. Notices to you go to your organization admins’ email addresses; notices to us go to legal@experientiallabs.ai.