FAQ
Built and tested, not yet deployed. Early access.
On this page
Is the endpoint live?
Not yet. The endpoint, the developer console and the MCP server are built and tested, but not deployed. When they open, you sign up in the console, make your own keys and get $5 of free credit.
Which model makes the decisions today?
A System One model for banking: one built to make one bounded decision in one step. It is pinned to one version, and a new version has to pass our published tests first. We plan to move to an open-weights System One model we can run ourselves. The model can change; the contract won’t.
What data does it need?
Most decisions ask only for a descriptor, a category code, amounts and dates. support_route, member_route and return_window also take text you choose to send, and the three health decisions take derived numbers or generic labels. Send no names, account numbers or balances. Bodies over 256 KB are refused. The three health decisions are built but not open, so do not send protected health information, including in your own questions.
What happens to our data?
For each answered call we log the decision name, the answer type, the abstain flag, the model, whether a rule settled it and how long the model took; for a failed call, the error’s type and status. Never your input. Every response is marked no-store. By default, the API requests routing only to model providers with zero-data-retention and no-training agreements; that isn’t yet verified by a live call. We’ll confirm the details in writing before your first live call.
Do you have SOC 2, a BAA or data residency?
We make no claims about SOC 2 reports, business associate agreements or data residency. If you need them, say so in the form and we’ll answer in writing. The three health decisions are built but not open. They take derived numbers and generic labels, plus one free-text member message that is not de-identified. We make no claim about HIPAA or a business associate agreement.
Can I use it for credit decisions?
No. Credit underwriting and adverse action carry their own notice rules under the Equal Credit Opportunity Act and Regulation B. The API doesn’t make those decisions, and its answers shouldn’t feed a credit line, an overdraft limit or an account closure. It doesn’t diagnose or verify identity either.
How fast and repeatable is it?
A full call adds one round trip to the decision model. That round trip, how repeatable the model’s answers are and whether its probabilities are calibrated aren’t measured yet.
What does it cost?
The planned price is $0.10 per million input tokens, with no charge for output. A new account gets $5 of free credit. Answers a rule settles without the model, calls that fail and replayed retries cost nothing. Nothing is charged until the API opens.
Is there an SDK?
Not yet. A TypeScript and a Python package open with early access. Until then any HTTP client works, and the types are ready to copy.
Is there a sandbox or test mode?
Not today.
Are there rate limits?
60 requests per minute per key. Over it, a 429 with Retry-After: 10. Limits
How is the API versioned?
The path carries /v1. A versioning and deprecation policy will be published before general availability.
Can I see which model answered?
Not in the response today.
Can it stop a card purchase before it goes through?
Whop decides each authorization itself and tells us afterwards, so the API acts ahead of time: freeze a card or set a per-transaction or period limit, which Whop applies to every later purchase, and check a purchase against those rules before it is tried.
Does it work with Apple Pay, Google Pay or a Swiss franc stablecoin?
Whop says its cards can be added to Apple Wallet and Google Wallet, so the API reports whether a card can be added and how; an Add to Wallet button inside our own app is not available. Whop holds USDT and accepts a list of dollar and euro stablecoins; francs are held as a fiat balance instead (chf_balance).
# FAQ Built and tested, not yet deployed. Early access. ## Is the endpoint live? Not yet. The endpoint, the developer console and the MCP server are built and tested, but not deployed. When they open, you sign up in the console, make your own keys and get $5 of free credit. ## Which model makes the decisions today? A System One model for banking: one built to make one bounded decision in one step. It is pinned to one version, and a new version has to pass our published tests first. We plan to move to an open-weights System One model we can run ourselves. The model can change; the contract won’t. ## What data does it need? Most decisions ask only for a descriptor, a category code, amounts and dates. support_route, member_route and return_window also take text you choose to send, and the three health decisions take derived numbers or generic labels. Send no names, account numbers or balances. Bodies over 256 KB are refused. The three health decisions are built but not open, so do not send protected health information, including in your own questions. ## What happens to our data? For each answered call we log the decision name, the answer type, the abstain flag, the model, whether a rule settled it and how long the model took; for a failed call, the error’s type and status. Never your input. Every response is marked no-store. By default, the API requests routing only to model providers with zero-data-retention and no-training agreements; that isn’t yet verified by a live call. We’ll confirm the details in writing before your first live call. ## Do you have SOC 2, a BAA or data residency? We make no claims about SOC 2 reports, business associate agreements or data residency. If you need them, say so in the form and we’ll answer in writing. The three health decisions are built but not open. They take derived numbers and generic labels, plus one free-text member message that is not de-identified. We make no claim about HIPAA or a business associate agreement. ## Can I use it for credit decisions? No. Credit underwriting and adverse action carry their own notice rules under the Equal Credit Opportunity Act and Regulation B. The API doesn’t make those decisions, and its answers shouldn’t feed a credit line, an overdraft limit or an account closure. It doesn’t diagnose or verify identity either. ## How fast and repeatable is it? A full call adds one round trip to the decision model. That round trip, how repeatable the model’s answers are and whether its probabilities are calibrated aren’t measured yet. ## What does it cost? The planned price is $0.10 per million input tokens, with no charge for output. A new account gets $5 of free credit. Answers a rule settles without the model, calls that fail and replayed retries cost nothing. Nothing is charged until the API opens. ## Is there an SDK? Not yet. A TypeScript and a Python package open with early access. Until then any HTTP client works, and the types are ready to copy. ## Is there a sandbox or test mode? Not today. ## Are there rate limits? 60 requests per minute per key. Over it, a 429 with `Retry-After: 10`. [Limits](/api/docs/limits) ## How is the API versioned? The path carries `/v1`. A versioning and deprecation policy will be published before general availability. ## Can I see which model answered? Not in the response today. ## Can it stop a card purchase before it goes through? Whop decides each authorization itself and tells us afterwards, so the API acts ahead of time: freeze a card or set a per-transaction or period limit, which Whop applies to every later purchase, and check a purchase against those rules before it is tried. ## Does it work with Apple Pay, Google Pay or a Swiss franc stablecoin? Whop says its cards can be added to Apple Wallet and Google Wallet, so the API reports whether a card can be added and how; an Add to Wallet button inside our own app is not available. Whop holds USDT and accepts a list of dollar and euro stablecoins; francs are held as a fiat balance instead (chf_balance). Source: /api/docs/faq