# ChatGPT Search Answers via OpenAI API: Citations & Sources (`crunchy_galaxy_qcr/chatgpt-search-answers`) Actor

Ask the model behind ChatGPT any list of questions with live web search, through OpenAI's official API, and get each answer with its citations, the pages it read and the searches it ran. No OpenAI account or key needed. Failed answers are never charged.

- **URL**: https://apify.com/crunchy_galaxy_qcr/chatgpt-search-answers.md
- **Developed by:** [Kendall Sorenson](https://apify.com/crunchy_galaxy_qcr) (community)
- **Categories:** AI, SEO tools
- **Features:** Limited permissions
- **Stats:** 2 total users, 1 monthly users, 100.0% runs succeeded, 0 bookmarks
- **User rating**: No ratings yet

## Pricing

$40.00 / 1,000 answers

This Actor is paid per event. You are not charged for the Apify platform usage, but only a fixed price for specific events.

Learn more: https://docs.apify.com/actors/running/actors-in-store.md#pay-per-event

## What's an Apify Actor?

An Actor is a serverless cloud program that runs on the Apify platform. It has two run modes.
In Batch mode, an Actor accepts a well-defined JSON input, performs an action which can take anything from a few seconds to a few hours,
and optionally produces a well-defined JSON output, datasets with results, or files in key-value store.
In Standby mode, an Actor provides a web server which can be used as a website, API, or an MCP server.

Apify vocabulary and the platform model are defined once, in the agent quickstart at https://apify.com/agents.md.

## How to integrate an Actor?

If asked about integration, you help developers integrate Actors into their projects.
You adapt to their stack and deliver integrations that are safe, well-documented, and production-ready.

Do not guess an integration path. Every one of them is in the agent quickstart at https://apify.com/agents.md: the Apify MCP server, Agent Skills with the Apify CLI, the JavaScript and Python clients, the REST API, and the account-free path for an agent with no human to sign in. It also carries the rule on stating cost before the first paid run.

For examples already wired to this Actor's own input schema, see the [API](#api) section below.

Each client library has reference documentation the quickstart does not restate: [JavaScript/TypeScript](https://docs.apify.com/api/client/js/docs.md) (`npm install apify-client`) and [Python](https://docs.apify.com/api/client/python/docs.md) (`pip install apify-client`).

# README

### ChatGPT Search Answers: the answer, its citations and its sources

Ask any list of questions and get back what the model behind ChatGPT answers after searching the web: the full answer, every page it cited, the pages it read without citing, and the searches it ran.

It uses **OpenAI's official API with live web search**. You don't need an OpenAI account or API key, and there are no cookies, logins or browser automation to break.

**Try it:** the form is filled with two prompts. That is 2 answers, **$0.08**.

![Real ChatGPT Search Answers output](https://api.apify.com/v2/key-value-stores/GMEuhx7aSQ9VxqkXd/records/chatgpt-search-answers-output.png?signature=AUTVBYM40XuNpiSOSBPq)

*A real answer from a run on November 11, 2026.*

**Failed answers are never charged.** You pay only for answers that were produced and saved.

#### What this is, and what it is not

This Actor calls OpenAI's API, using a current GPT model with the web search tool. It is the closest official equivalent of a ChatGPT search answer, and it is stable and repeatable.

It is **not** a recording of the ChatGPT app. Answers in the app can differ, because the app adds its own instructions, memory and personalisation. If you need to know how AI models answer a question and which sites they rely on, this gives you that reliably. If you need a screenshot of exactly what one person saw in the app, this is not that.

#### What you get for each answer

- **The full answer** in Markdown
- **Citations**: each page cited in the answer, in order, with URL, title and website. Tracking parameters are removed so the same page counts once.
- **Pages read but not cited**: what the model opened during its search and left out
- **The searches it ran**: the actual queries the model sent to the web
- The exact model version, country and time

#### A summary of the whole run

Each run also adds one **free** summary row: how many answers cited at least one page, the average number of citations, and the **websites cited by the most answers**. Ask 50 questions about your market and you can see at a glance which sites the model leans on.

#### Ask the same question more than once

AI answers vary from one request to the next. Set **Answers per prompt** to 3 or 5 to see how stable an answer is and which sources keep coming back.

#### Input

| Field | What it does |
|---|---|
| Prompts | One question per line, up to 1,000 characters each |
| Country for web search | Two-letter code such as `US`, `GB` or `DE`. The search behaves as if the user were there |
| Answer language | Optional, for example `German` |
| Answers per prompt | 1 to 10 |
| Always search the web | On by default, so every answer can cite sources |
| Search depth | Low, medium or high. Same price |
| Include pages read but not cited | On by default |
| Maximum answers | Stops the run after this many |

Example input:

```json
{
    "prompts": ["What is the best CRM for a small marketing agency?", "Which standing desks do reviewers recommend in 2026?"],
    "country": "US",
    "samplesPerPrompt": 1
}
```

#### Output

A real answer row from a run on November 11, 2026, with the answer text and lists shortened:

```json
{
    "type": "answer",
    "prompt": "Which standing desks do reviewers recommend in 2026?",
    "sample": 1,
    "answerText": "In 2026, reviewers most often recommended these standing desks: …",
    "usedWebSearch": true,
    "searchQueries": ["best standing desks 2026 review standing desk recommendations"],
    "citationCount": 5,
    "citations": [
        {
            "position": 1,
            "url": "https://www.tomsguide.com/home/home-office/best-standing-desks",
            "title": "I've tested more than 27 standing desks and these are the best ones for your home office",
            "domain": "tomsguide.com"
        }
    ],
    "citedDomains": ["tomsguide.com"],
    "consultedSources": [{ "url": "https://www.tomsguide.com/home/home-office", "title": null, "domain": "tomsguide.com" }],
    "model": "gpt-5.4-mini-2026-03-17",
    "country": "US",
    "checkedAt": "2026-11-12T01:18:34.764Z"
}
```

Results export to JSON, CSV or Excel.

#### How much does it cost?

You pay per answer, with no monthly fee:

| Event | Price |
|---|---|
| Answer (with citations, sources and searches) | $0.04 |
| Run summary row | Free |
| Answers that fail | Free |

Worked examples:

- 25 questions, one answer each: **$1.00**
- 50 questions, 3 answers each: **$6.00**

The price covers the AI model and the web search for each answer. Search depth does not change it.

#### Apify Free plan: 3 answers per run

Every answer costs real money to produce, so accounts on the Apify Free plan are limited to **3 answers per run** as a trial. Runs on any paid Apify plan have no trial limit. If a run is cut short by this limit, the log and status message say so, and you are charged only for the answers you received.

#### Use it from AI agents (MCP)

The Actor works as a tool for AI assistants through Apify's MCP server:

```
https://mcp.apify.com?tools=crunchy_galaxy_qcr/chatgpt-search-answers
```

The only field an agent has to set is `prompts`. Every output field has a description, and citations come back as structured URL, title and domain, ready to check or compare.

#### Code example

```javascript
import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: '<YOUR_API_TOKEN>' });
const run = await client.actor('crunchy_galaxy_qcr/chatgpt-search-answers').call({
    prompts: ['What is the best CRM for a small marketing agency?'],
    country: 'US',
    samplesPerPrompt: 3,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
for (const row of items.filter((r) => r.type === 'answer')) console.log(row.citedDomains, row.answerText.slice(0, 200));
```

#### FAQ

##### Which model answers the questions?

A current small GPT model from OpenAI, chosen automatically, with OpenAI's hosted web search tool. The exact model version is recorded in every row.

##### Is this the same answer I would get in the ChatGPT app?

Not exactly. See "What this is, and what it is not" above. The API answer comes from the same family of models with web search, but the app can add personal context that the API does not have.

##### Why do two runs of the same prompt give different answers?

AI models don't give identical answers every time, and the web changes. Use **Answers per prompt** to see the spread instead of trusting one answer.

##### Are the answers always correct?

No. These are AI-generated answers and can be wrong or out of date. The citations are there so you can check them.

##### Do you store my prompts?

Prompts are sent to OpenAI to produce the answers, with OpenAI's storage of responses turned off for these requests. Your prompts and results stay in your own Apify account like any other run.

##### What can't I use this for?

Don't send prompts that break OpenAI's usage policies. This Actor is for questions and research, not for generating harmful content.

##### Want brand tracking instead of raw answers?

If your goal is to know whether AI recommends your brand, or which sites it cites in your market over time, the two trackers below do that analysis for you.

#### Related tools

- [AI Visibility Tracker](https://apify.com/crunchy_galaxy_qcr/ai-visibility-tracker): whether ChatGPT, Gemini and Perplexity recommend your brand, with rank and share of voice.
- [AI Citation Tracker](https://apify.com/crunchy_galaxy_qcr/ai-citation-tracker): which websites AI answers cite in your market, and where competitors are cited instead of you.
- [Google Autocomplete Scraper](https://apify.com/crunchy_galaxy_qcr/google-autocomplete-keyword-ideas): find the questions people actually ask, then run them through this Actor.

Found a problem or want a field added? Open an Issue on this Actor's page.

# Actor input Schema

## `prompts` (type: `array`):

One question or prompt per line, up to 1,000 characters each. Each one is answered separately, with a fresh web search.

## `country` (type: `string`):

Two-letter country code, for example US, GB or DE. The web search behaves as if the user were in this country. Leave empty for no location.

## `language` (type: `string`):

Optional. For example English, German or Spanish. Leave empty to let the model answer in the prompt's language.

## `samplesPerPrompt` (type: `integer`):

AI answers vary from one request to the next. Ask the same prompt several times to see how much. Each answer is charged.

## `forceWebSearch` (type: `boolean`):

On: every answer uses a live web search and can cite sources. Off: the model decides, and may answer from memory with no citations.

## `searchDepth` (type: `string`):

How much web content the model reads before answering. Higher reads more pages and takes longer. The price per answer is the same.

## `includeConsultedSources` (type: `boolean`):

Add the pages the model opened during its search but did not cite in the answer.

## `maxAnswers` (type: `integer`):

Stops the run after this many answers. Apify Free plan accounts are limited to 3 answers per run as a trial; paid plans have no trial limit.

## Actor input object example

```json
{
  "prompts": [
    "What is the best CRM for a small marketing agency?",
    "Which standing desks do reviewers recommend in 2026?"
  ],
  "country": "US",
  "samplesPerPrompt": 1,
  "forceWebSearch": true,
  "searchDepth": "medium",
  "includeConsultedSources": true,
  "maxAnswers": 100
}
```

# Actor output Schema

## `answers` (type: `string`):

Table of answers: prompt, answer text, number of citations, cited sites and the searches the model ran.

## `allResults` (type: `string`):

Every row with every field, including each citation's URL and title, the pages read but not cited, and the run summary with the most-cited sites.

## `summary` (type: `string`):

Counts for this run: answers saved, answers that failed (not charged) and the model used.

# API

You can run this Actor programmatically using our API. Below are code examples in JavaScript, Python, and CLI, as well as the OpenAPI specification and MCP server setup.

## JavaScript example

```javascript
import { ApifyClient } from 'apify-client';

// Initialize the ApifyClient with your Apify API token
// Replace the '<YOUR_API_TOKEN>' with your token
const client = new ApifyClient({
    token: '<YOUR_API_TOKEN>',
});

// Prepare Actor input
const input = {
    "prompts": [
        "What is the best CRM for a small marketing agency?",
        "Which standing desks do reviewers recommend in 2026?"
    ],
    "country": "US"
};

// Run the Actor and wait for it to finish
const run = await client.actor("crunchy_galaxy_qcr/chatgpt-search-answers").call(input);

// Fetch and print Actor results from the run's dataset (if any)
console.log('Results from dataset');
console.log(`💾 Check your data here: https://console.apify.com/storage/datasets/${run.defaultDatasetId}`);
const { items } = await client.dataset(run.defaultDatasetId).listItems();
items.forEach((item) => {
    console.dir(item);
});

// 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/js/docs

```

## Python example

```python
from apify_client import ApifyClient

# Initialize the ApifyClient with your Apify API token
# Replace '<YOUR_API_TOKEN>' with your token.
client = ApifyClient("<YOUR_API_TOKEN>")

# Prepare the Actor input
run_input = {
    "prompts": [
        "What is the best CRM for a small marketing agency?",
        "Which standing desks do reviewers recommend in 2026?",
    ],
    "country": "US",
}

# Run the Actor and wait for it to finish
run = client.actor("crunchy_galaxy_qcr/chatgpt-search-answers").call(run_input=run_input)

# Fetch and print Actor results from the run's dataset (if there are any)
print(f"💾 Check your data here: https://console.apify.com/storage/datasets/{run.default_dataset_id}")
for item in client.dataset(run.default_dataset_id).iterate_items():
    print(item)

# 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/python/docs/quick-start

```

## CLI example

```bash
echo '{
  "prompts": [
    "What is the best CRM for a small marketing agency?",
    "Which standing desks do reviewers recommend in 2026?"
  ],
  "country": "US"
}' |
apify call crunchy_galaxy_qcr/chatgpt-search-answers --silent --output-dataset

```

## MCP server setup

```json
{
    "mcpServers": {
        "apify": {
            "type": "http",
            "url": "https://mcp.apify.com/?tools=fetch-actor-details,crunchy_galaxy_qcr/chatgpt-search-answers"
        }
    }
}
```

The hosted server signs you in with OAuth on first connect, so no API token belongs in this config. Clients without OAuth support can send an `Authorization: Bearer <APIFY_API_TOKEN>` header instead, using a token from API & Integrations in Apify Console (https://console.apify.com/settings/integrations).

## OpenAPI specification

Download the OpenAPI definition: https://api.apify.com/v2/actors/JifCNePG5hZlWrHfJ/builds/IEmBNgvxJH8M9gDmX/openapi.json
