> ## Documentation Index
> Fetch the complete documentation index at: https://docs.scrapebadger.com/llms.txt
> Use this file to discover all available pages before exploring further.

# ChatGPT API Overview

> Ask the real chatgpt.com a question and get the answer back as structured JSON — with the web sources it cited, anchored to the exact text they support. No OpenAI account or API key required.

## Overview

The ScrapeBadger **ChatGPT API** sends a prompt to **the real `chatgpt.com`** —
not the OpenAI API — and returns the answer as structured JSON, including the
web sources ChatGPT actually browsed and cited.

Requests are **anonymous**: no OpenAI account, no OpenAI API key, no billing
relationship with OpenAI. You see what a logged-out user in your chosen country
sees.

<Info>
  All endpoints live under `https://scrapebadger.com/v1/chatgpt/*` and
  authenticate with the `X-API-Key` header. `/ask` and `/brand-visibility`
  accept both `GET` (query params) and `POST` (JSON body). Credits are charged
  per request and reported on the `X-Credits-Used` response header.
</Info>

<Warning>
  **Answers take 20-25s ungrounded, 30-70s with web search.** This is a live browse of chatgpt.com, not a
  cached lookup — the clock includes ChatGPT composing the answer and, when it
  browses, running its own web search. Set your HTTP client timeout to **60 s
  or more**; the default 5–10 s timeout in most SDKs will abort the call.
</Warning>

## Features

* **Real ChatGPT answers** — the full text as rendered on chatgpt.com, plus a
  Markdown variant when one is available.
* **Anchored citations** — every citation carries `start_index` / `end_index`
  character offsets into `answer` and the `matched_text` it supports, so you can
  highlight exactly which sentence a source backs.
* **The full search trail** — `search_results` is everything ChatGPT retrieved,
  each flagged `cited: true/false`. `citations` is the subset it actually
  referenced. Most APIs only give you the second list.
* **Did it really browse?** — `web_search_triggered` tells you whether ChatGPT
  ran a web search or answered from its own weights. Force it with
  `web_search=force`, or suppress it with `web_search=off`.
* **Brand visibility (AEO/GEO) built in** — one call returns mention counts,
  position score, share of voice against named competitors, and whether your
  domain was cited. No separate subscription.
* **Country control** — `country` picks the egress country, so you can compare
  the answer a user in `DE` gets against the one in `US`.

## Honesty caveat

<Warning>
  Each request is a **single, fresh, anonymous question**. There is no
  conversation memory, no multi-turn follow-up, no model selection, no
  streaming, and no image input — `conversation_id` and `message_id` identify
  the one-shot exchange, they are not handles you can continue from. The
  `model` field reports which model answered; it is **not** a request
  parameter, and `/models` is a read-only catalogue of what chatgpt.com
  currently offers.

  ChatGPT is non-deterministic: the same prompt asked twice can produce
  different wording and different sources. For monitoring use cases, sample
  repeatedly and aggregate rather than trusting a single call.
</Warning>

## Credits

| Endpoint         | Path                                        | Credits |
| ---------------- | ------------------------------------------- | ------- |
| Ask              | `GET`/`POST` `/v1/chatgpt/ask`              | 20      |
| Brand visibility | `GET`/`POST` `/v1/chatgpt/brand-visibility` | 25      |
| List models      | `GET /v1/chatgpt/models`                    | 1       |
| Health           | `GET`/`HEAD` `/v1/chatgpt/health`           | 0       |

<Note>
  At the PAYG rate of $0.15 per 1K credits, an `/ask` call is **$0.003\*\* —
  **\$3.00 per 1,000 answers**. Subscription tiers cut that substantially. Non-2xx
  responses are never billed.
</Note>

## Quickstart

<CodeGroup>
  ```bash cURL theme={null}
  curl "https://scrapebadger.com/v1/chatgpt/ask?prompt=What%20are%20the%20best%20web%20scraping%20APIs%20in%202026%3F" \
    -H "X-API-Key: YOUR_API_KEY" \
    --max-time 60
  ```

  ```javascript Node.js theme={null}
  const res = await fetch("https://scrapebadger.com/v1/chatgpt/ask", {
    method: "POST",
    headers: {
      "X-API-Key": process.env.SCRAPEBADGER_API_KEY,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      prompt: "What are the best web scraping APIs in 2026?",
      country: "US",
      web_search: "force",
    }),
    signal: AbortSignal.timeout(60_000),
  });
  const data = await res.json();
  console.log(data.answer);
  console.log(data.source_domains);
  ```

  ```python Python theme={null}
  import requests

  res = requests.post(
      "https://scrapebadger.com/v1/chatgpt/ask",
      headers={"X-API-Key": "YOUR_API_KEY"},
      json={
          "prompt": "What are the best web scraping APIs in 2026?",
          "country": "US",
          "web_search": "force",
      },
      timeout=60,
  )
  data = res.json()
  print(data["answer"])
  print(data["source_domains"])
  ```
</CodeGroup>

## What you can build

<CardGroup cols={2}>
  <Card title="AEO / GEO monitoring" icon="chart-line">
    Track whether ChatGPT names your brand for the prompts your buyers ask, how
    early it names you, and how you rank against competitors — with
    [`/brand-visibility`](/api-reference/endpoint/chatgpt/brand-visibility).
  </Card>

  <Card title="Citation-source auditing" icon="link">
    Find which domains ChatGPT trusts for a topic. `search_results` shows
    everything it retrieved; `cited` shows what survived into the answer.
  </Card>

  <Card title="Answer-quality research" icon="magnifying-glass">
    Compare answers across countries, or with `web_search` forced on and off, to
    separate what ChatGPT knows from what it looked up.
  </Card>

  <Card title="Grounded content pipelines" icon="quote-left">
    Use `start_index` / `end_index` to attach a real source to each claim before
    republishing anything ChatGPT wrote.
  </Card>
</CardGroup>

## How we compare

| Vendor                 | Price / 1K         | Citations              | Full search trail | Brand analysis   |
| ---------------------- | ------------------ | ---------------------- | ----------------- | ---------------- |
| Bright Data            | \$1.50             | Yes                    | No                | No               |
| Apify `chatgpt-search` | \$3.00             | Yes                    | No                | No               |
| DataForSEO LLM Scraper | \$1.20–4.00        | Yes                    | Yes               | Separate product |
| Scrapingdog            | \~30 credits/req   | **No**                 | No                | No               |
| ScrapingBee            | Credits            | **No**                 | No                | No               |
| **ScrapeBadger**       | **\$3.00** (20 cr) | **Yes + text offsets** | **Yes**           | **Bundled**      |

Five things that are genuinely ours:

1. **Citations anchored to the answer.** `start_index` / `end_index` /
   `matched_text` tell you *which sentence* a source supports. No other vendor
   returns text offsets.
2. **`web_search_triggered`.** Only Bright Data otherwise tells you whether
   ChatGPT really browsed — everyone else leaves you guessing whether a
   citation-free answer came from memory or a failed search.
3. **`search_results` vs `citations`.** The retrieved set *and* the referenced
   subset. Only DataForSEO separates these.
4. **Brand visibility in the same API.** Profound, Otterly, Peec and AthenaHQ
   sell this as a \$99–989/month subscription. Here it is a 25-credit call.
5. **Published typically under 30s ungrounded live latency.** DataForSEO's live tier quotes a 90 s
   SLA.

## FAQ

<AccordionGroup>
  <Accordion title="Is this the OpenAI API?">
    No. This drives the public `chatgpt.com` web interface as an anonymous
    visitor. You do not need an OpenAI account, an OpenAI API key, or an OpenAI
    subscription — and answers include web citations, which the OpenAI
    completions API does not return.
  </Accordion>

  <Accordion title="Why does a call take 20-25s ungrounded, 30-70s with web search?">
    Because a real answer is being generated, and when ChatGPT browses it runs
    its own web search first. Budget a 60 s client timeout and run calls
    concurrently rather than in a tight sequence.
  </Accordion>

  <Accordion title="Can I continue a conversation or ask a follow-up?">
    No. Every request is a fresh, standalone question with no memory of previous
    calls. Put all necessary context in the single `prompt` (up to 4,096
    characters).
  </Accordion>

  <Accordion title="Can I choose which model answers?">
    No. `/models` lists the model slugs chatgpt.com currently exposes, and the
    `model` field in a response reports which one answered — but you cannot
    request a specific one.
  </Accordion>

  <Accordion title="I got an answer with no citations. Is that a bug?">
    Check `web_search_triggered`. If it is `false`, ChatGPT answered from its
    training data and there were no sources to cite — that is normal for
    general-knowledge prompts. Set `web_search=force` if you always want it to
    browse.
  </Accordion>

  <Accordion title="Will the same prompt always give the same answer?">
    No. ChatGPT is non-deterministic, and the live web changes. For tracking,
    sample the same prompt on a schedule and look at the trend, not a single
    result.
  </Accordion>

  <Accordion title="Does country change the answer?">
    It changes the egress country of the request, which affects the localised
    results ChatGPT retrieves and can change the answer's emphasis and language.
    Pass any ISO-3166 alpha-2 code; the default is `US`.
  </Accordion>

  <Accordion title="What happens if a request fails?">
    You get `{"detail": {"code": ..., "message": ...}}` with an HTTP status of
    422, 429 or 502 depending on the cause. **Non-2xx responses are never
    billed**, so a retry costs you nothing extra beyond the successful call.
  </Accordion>
</AccordionGroup>

<Tip>
  Start with [`/ask`](/api-reference/endpoint/chatgpt/ask) to see the shape of a
  real answer, then move to
  [`/brand-visibility`](/api-reference/endpoint/chatgpt/brand-visibility) once
  you know which prompts matter for your market.
</Tip>