Ask ChatGPT
TL;DR
About
A growth lead at a French payroll startup sends "What payroll software should a 20-person company in France use?" from a signed-out browser every Monday and logs which vendors the reply names and which review sites it cites, since those are the pages worth getting listed on. Signed out is the closest thing to a stranger asking, but you get little beyond the reply text and its cited references. A signed-in account adds the model id, every query ChatGPT typed and every page those searches pulled up, though saved memory can shape the reply and the chat is kept in that account's history. Those queries are worth a close read, since one question can fan out into many (a September run typed 9 across 2 search calls). Each run is one question in a fresh chat, and setting mode to work is saved on the account, so your own next chat starts in Work too.
Input
question requiredThe question to send as the first message of a new chat.
Output
calls requiredOne entry per search call. A call often runs a whole round of queries at once and answers one list for all of them. What the page does not receive: which query returned which result, the call's arguments, and which search backend answered it. Empty when it did not search. Null when this run could not see the searches (signed out).
model requiredWhich model answered, by ChatGPT's own id (e.g. "gpt-6"). Null when the app does not say.
answer requiredThe last entry of `answers`. Null signed out: `answers` has its text there.
answers requiredOne entry per reply, usually one; ChatGPT sometimes answers twice and asks which reply you prefer, and neither entry is more official than the other — read the length before reading answers[0]. Signed in, each is ChatGPT's own markdown of the reply, with each citation as a markdown link to its source and widgets (images, forms) left out. Signed out, the reply's text as rendered.
sources requiredEvery page the searches put in front of the model, once each, in the order it got them — the pool the answer could draw on, as opposed to `references`, what it cited. Empty when it did not search. Null when this run could not see the searches (signed out).
searches requiredOne entry per round of web searches, in order: what the model typed and the results put in front of it. Empty when it did not search. Null when this run could not see the searches (signed out).
references requiredEvery source the answer CITED or linked — not everything the search retrieved (that is `sources`) — deduplicated by url and in the order the answer surfaced them. Empty when the answer cited nothing, which is a fact about that reply: the same question searches on one run and not the next. Null when this run could not see the citations (signed out): unknown, not none.
stopReason requiredWhy generation ended, as ChatGPT records it: "stop" for an answer the model finished. Null when ChatGPT records no reason, and signed out. A finished answer can still be only a sentence announcing work it never did; that is what the model returned, not a truncation.
searchContext requiredWhat the server says about this turn, as sent to the page. Every field is null signed out, and a field is null when the app did not send it.
conversationId requiredThe conversation's id. Null if the app did not create one.
webSearchQueries requiredThe searches ChatGPT issued, in order. Empty when the reply did not search. Null when this run could not see the searches (signed out).
Example output
Shape only: placeholder values generated from the output schema, not a real run.
{
"calls": [
{
"call": 3,
"tool": "…",
"results": [
{
"ref": "…",
"url": "https://example.com/item/123",
"rank": 3,
"type": "…",
"title": "Example",
"pubDate": "2026-01-15T09:30:00Z",
"snippet": "…",
"attribution": "…",
"thumbnailUrl": "https://example.com/item/123"
}
]
}
],
"model": "…",
"answer": "…",
"answers": [
"…"
],
"sources": [
{
"url": "https://example.com/item/123",
"title": "Example",
"attribution": "…"
}
],
"searches": [
{
"queries": [
"…"
],
"results": [
{
"ref": "…",
"url": "https://example.com/item/123",
"type": "…",
"title": "Example",
"pubDate": "2026-01-15T09:30:00Z",
"attribution": "…"
}
]
}
],
"references": [
{
"url": "https://example.com/item/123",
"title": "Example",
"attribution": "…"
}
],
"stopReason": "…",
"searchContext": {
"tool": "…",
"useCase": "…",
"toolInvoked": true,
"locationUsed": "…",
"clusterRegion": "…",
"locationIsPrecise": true
},
"conversationId": "abc123",
"webSearchQueries": [
"…"
]
}FAQ
ChatGPT billing: list your ChatGPT subscription invoices
ChatGPT subscription invoices, not the OpenAI API platform (that is platform.openai.com/list_invoices). List the billing history of every ChatGPT account the signed-in user belongs to — the personal account and each Business, Team or Enterprise workspace — whichever one the browser currently has open. Each invoice comes with its date, amount, currency, payment status, the plan it paid for, a link to the receipt, and the account it belongs to; a separate accounts list shows each account, how many invoices it has, and whether its billing could be read. The full history is returned, not only the most recent entries. Amounts come back in the currency's smallest unit, so 1917 means 19.17 EUR. When the browser is not signed in to ChatGPT, the result says so with signed_in set to false and empty lists; check it before reading an empty list as an account that was never billed. Feed hosted_invoice_url to invoice.stripe.com/download_invoice_pdf to get the PDF.
Who am I (ChatGPT)
Report which ChatGPT account the browser is signed into: id, name, email, picture, identity provider — or loggedIn:false when the browser is anonymous. Reads the session endpoint the app itself hydrates auth from; never returns the session's access token.
OpenAI (ChatGPT) login (Google SSO)
Log into OpenAI / ChatGPT via Google SSO — picks the right account when several are signed in, and does nothing if you're already logged in. If Google interrupts the flow with a password re-confirmation or a 2FA step-up, the script can't complete that step on its own; sign in once manually in the same browser to clear the step-up, then re-run. It may also occasionally fail to complete a fresh login due to the site's own anti-automation checks — if that happens, try again shortly.