PUBLIC API
Strict, transient payment-format QA endpoints
Base URL: https://bin.chatgpt.org.uk/api/v1. Send test data only. Never submit real cardholder information.
Limits and behavior
- JSON only; numbers must be strings.
- Maximum request body: 8 KiB.
- Maximum API batch: 20 values.
- 120 requests per minute per client and route, enforced at the Cloudflare location.
- Responses use
Cache-Control: no-storeand include a request ID.
POST /validate
curl -X POST 'https://bin.chatgpt.org.uk/api/v1/validate' \
-H 'Content-Type: application/json' \
-d '{"numbers":["4111111111111111","0"]}'Returns valid, luhnValid, shapeValid, masked output, reason codes, and overlap-aware network hints. Read networkCandidates as the complete candidate list and networkAmbiguous as the overlap flag. networkId remains the first matching candidate for backward compatibility, so it must not be treated as the only network when networkAmbiguous is true.
POST /generate
curl -X POST 'https://bin.chatgpt.org.uk/api/v1/generate' \
-H 'Content-Type: application/json' \
-d '{"prefix":"411111","count":5,"networkId":"visa","length":16,"mode":"valid","seed":"regression-42"}'Modes: valid, invalid-luhn, and invalid-length. API generation is capped at 20; the local workbench supports 100.
POST /suite
curl -X POST 'https://bin.chatgpt.org.uk/api/v1/suite' \
-H 'Content-Type: application/json' \
-d '{"seed":"checkout-regression"}'Returns a fixed 21-case matrix across seven common card networks and three isolated format states. Optional modes can select a unique subset of valid, invalid-luhn, and invalid-length.
GET /presets and /health
/presets returns the fixed scenario catalog with authoritative documentation links. /health returns the service version.
Error model
{
"success": false,
"error": {
"code": "NUMBER_TYPE",
"message": "Every card number must be a string to prevent precision loss."
},
"requestId": "..."
}