ChatGPT Search Scraper: Answers, Citations, Brand Tracking
Pricing
from $9.00 / 1,000 answers
ChatGPT Search Scraper: Answers, Citations, Brand Tracking
Ask ChatGPT your prompts in a fresh logged out chat and get one row per answer: the full text, every cited source with its position, the brands and products named, and whether each brand you track was mentioned, cited or linked. Track AI visibility. No API key.
Pricing
from $9.00 / 1,000 answers
Rating
0.0
(0)
Developer
The Mine Works
Maintained by CommunityActor stats
0
Bookmarked
1
Total users
0
Monthly active users
10 days ago
Last modified
Categories
Features
Share
ChatGPT Search Scraper: Answers, Citations and Brand Tracking
Part of The Mine Works Marketing, SEO and reviews family: Facebook Ad Library Scraper, Google Ads Transparency Scraper, Similarweb Scraper, Google News Scraper, Trustpilot Reviews Scraper, Google Trends Scraper.
Ask ChatGPT the questions your customers ask and get back what a visitor sees: the full answer, every source it linked or cited, the brands and products it named, and whether the brands you track were mentioned, cited or linked. One row per prompt. No OpenAI account, no API key, no login.
It is built for AI visibility work: finding out whether ChatGPT recommends you or a competitor for the questions that matter, which sites it trusts, and how that changes week by week.
✅ No account or API key | ✅ Fresh logged out chat for every prompt | ✅ Cited sources with position | ✅ Brand tracking built in | ✅ Pay only for answers delivered | ✅ MCP ready for AI agents
Why choose this actor?
- It shows what a real visitor gets. Each prompt is typed into a new logged out chat on chatgpt.com, so no history, memory or custom instructions change the answer.
- Brand tracking in every row. Give your brand and your competitors once. Each answer says whether each brand was named, how often, where it first appears, its rank among your brands, and whether its own site was cited or linked.
- Sources with their position. Every linked and cited page comes back with its title, URL, domain and order, so you can see which sites ChatGPT leans on.
- Answers from another country. ChatGPT tailors shops, prices and brands to where the visitor is. Pick a country and the question is asked from there.
- You pay per answer. Failed prompts and the summary row are never charged.
Try it in one minute
- Open the actor and keep the prefilled prompt,
best crm software for small business. - Keep the three prefilled brands: HubSpot, Zoho and Salesforce.
- Press Start. One answer row appears in your dataset in under a minute.
Apify's free plan includes $5 of credit every month, which covers about 330 answers at this actor's price.
Copy to your AI assistant
Copy this and paste it into Claude, ChatGPT or Cursor to call the actor:
themineworks/chatgpt-search-scraper on Apify. Asks ChatGPT prompts in a fresh logged out chat and returns one row per answer: answer_text, answer_markdown, sources (position, title, url, domain), cited_domains, entities and tracked_brands (mentioned, mention_count, first_position, rank_among_tracked, cited, linked). Input: prompts (array of questions, up to 500), brandsToTrack (array, optional; use "Zoho|Zoho CRM|zoho.com" to join spellings), country (two letter code, default "US"), webSearch (boolean, default true), maxConcurrency (1 to 5, default 2). Call ApifyClient("TOKEN").actor("themineworks/chatgpt-search-scraper").call(run_input={"prompts": ["best crm software for small business"]}), then client.dataset(run["defaultDatasetId"]).list_items().items. Pay per event: answer $15.00 per 1,000 on the Free plan (lower on paid plans), plus $0.003 per run. Other countries are charged at the local answer price. Full spec: GET https://api.apify.com/v2/acts/themineworks~chatgpt-search-scraper/builds/default (Bearer TOKEN). Token: https://console.apify.com/account/integrations?fpr=ymnoit&utm_source=apify-readme&utm_medium=referral
Key features
- Prompts: up to 500 per run, each asked in its own new chat.
- Brands to track: your brand and competitors, with alias spellings and domains.
- Web search switch: ask ChatGPT with its Web search tool on, as a visitor would, or let it decide.
- 24 countries: US is answered directly. Every other country goes through a residential connection in that country.
- Run summary: a free summary row and an
OUTPUTrecord report prompts asked, answered and failed. - API, schedules and MCP: call it from the API, run it weekly on a schedule, or let an AI agent use it through the Apify MCP server.
How to use it
Basic: one question
{"prompts": ["best crm software for small business"]}
Track your brand against competitors
{"prompts": ["best crm software for small business","hubspot alternatives","which crm is easiest to set up"],"brandsToTrack": ["HubSpot", "Zoho|Zoho CRM|zoho.com", "Salesforce"]}
Join spellings and the website with |. Every row then carries a tracked_brands list in the same order as yours.
Ask from another country
{"prompts": ["best running shoes under 100"],"country": "GB"}
Prices, shops and brands in the answer follow the country. Anything other than US is charged at the local answer price.
Watch it over time
Run the same prompts and brands every week with an Apify schedule. Each row has a prompt_id that stays the same for the same prompt and country, so you can chart any_brand_mentioned and rank_among_tracked by run_date.
Input parameters
| Field | Type | Default | What it does |
|---|---|---|---|
prompts | array of strings | prefilled | Questions to ask, one per line, up to 500 per run. |
brandsToTrack | array of strings | none | Brands to look for in each answer. Join spellings with ` |
country | string | US | Where the question is asked from. One of 24 countries. |
webSearch | boolean | true | Switch on ChatGPT's Web search tool in the chat. |
maxConcurrency | integer | 2 | Prompts asked at the same time (1 to 5; two works best, more tabs slow each other down). |
What data do you get?
One row per prompt. The main fields:
| Field | Meaning |
|---|---|
prompt, prompt_id, country, run_date | What was asked, a stable id for it, where from and when. |
answer_text, answer_markdown | The full answer as plain text and with its Markdown (tables and lists kept). |
answer_word_count | Length of the answer in words. |
web_search_used | Whether ChatGPT searched the web for this answer. |
sources | Linked and cited pages: position, title, url, domain, publisher, section, times_cited. |
source_count, cited_source_count, cited_domains, source_urls | Totals and the domains and URLs behind the answer. |
entities, entity_count | Brands, products and places ChatGPT named, with a category. |
tracked_brands | One entry per brand you gave: mentioned, mention_count, first_position, first_sentence, rank_among_tracked, cited, cited_url, linked, linked_url. |
brands_mentioned, brands_cited, any_brand_mentioned, any_brand_cited | Quick flags for filters and charts. |
response_seconds, answered_at, conversation_id | Timing and the chat id. |
Output example
{"prompt": "best crm software for small business","country": "US","run_date": "2026-11-10","answer_text": "For most small businesses, I recommend HubSpot, Zoho CRM, or Pipedrive. The best choice depends on...","answer_word_count": 527,"web_search_used": true,"sources": [{ "position": 1, "title": "Visit HubSpot", "url": "https://www.hubspot.com/products/crm", "domain": "hubspot.com", "section": "linked" },{ "position": 2, "title": "Visit Zoho CRM", "url": "https://www.zoho.com/crm/", "domain": "zoho.com", "section": "linked" }],"source_count": 11,"cited_source_count": 4,"cited_domains": ["techradar.com", "g2.com", "unpackmoney.com", "spotsaas.com"],"entities": [{ "name": "HubSpot CRM", "category": "software" },{ "name": "Zoho CRM", "category": "software" }],"tracked_brands": [{ "brand": "HubSpot", "mentioned": true, "mention_count": 4, "first_position": 40, "rank_among_tracked": 1, "cited": false, "linked": true, "linked_url": "https://www.hubspot.com/products/crm" },{ "brand": "Zoho", "mentioned": true, "mention_count": 7, "first_position": 49, "rank_among_tracked": 2, "cited": false, "linked": true, "linked_url": "https://www.zoho.com/crm/" }],"brands_mentioned": ["HubSpot", "Zoho", "Salesforce"],"any_brand_mentioned": true,"any_brand_cited": false,"response_seconds": 59.9,"answered_at": "2026-11-10T16:20:24.090Z"}
Pricing
Pay per event: you pay for the answers you receive, plus a flat $0.003 per run, whatever memory you choose.
| Event | When it is charged | Free plan | Bronze (Starter) | Silver (Scale) | Gold (Business) |
|---|---|---|---|---|---|
chatgpt-answer | Once per answer row for US | $15.00 per 1,000 | $12.50 per 1,000 | $11.00 per 1,000 | $9.00 per 1,000 |
local-country-answer | Once per answer row for any other country | $95.00 per 1,000 | $85.00 per 1,000 | $80.00 per 1,000 | $75.00 per 1,000 |
run-start | Once per run, after the first answer arrived | $0.003 | $0.003 | $0.003 | $0.003 |
The price of one answer is the price per 1,000 divided by 1,000. Other countries cost more because each answer travels through a residential connection in that country.
Never charged:
- prompts that ChatGPT did not answer, and the free summary row;
- a prompt repeated in the same run.
Run it on a schedule
Open the actor, press Schedule, and pick weekly or daily with the same input. Each run adds rows with a new run_date. Compare any_brand_mentioned and rank_among_tracked over time, or export the dataset to Google Sheets or your BI tool.
FAQ
What does this actor read?
The answer page of chatgpt.com for a logged out visitor, the same text and sources you would see in a private browser window.
Is this the same as the OpenAI API?
No. The API answers without ChatGPT's web search and shopping layer. This actor returns what the ChatGPT website shows a visitor, including linked and cited pages. It is the better guide to what people actually see.
Why do some answers have no sources?
ChatGPT decides whether a question needs the web. For plain knowledge questions, such as how photosynthesis works, it often answers without searching and there are no sources. Questions about products, services, prices and recommendations nearly always have sources. The web_search_used field tells you which case you got.
Will I get the same answer twice?
Not always. ChatGPT answers vary from one chat to the next, so run important prompts more than once and look at the pattern, not a single answer.
How long does a run take?
A browser has to start first, then each answer takes about 15 to 30 seconds with two prompts at the same time. Ten prompts took about two minutes in our test. The actor needs 4 GB of memory and sets this itself.
Which models and plan does it see?
The logged out guest experience ChatGPT shows at the time of the run. The model_experience and plan fields record what was served.
Do I need an account, cookies or a proxy?
No. For the US the actor connects directly. For other countries it uses a residential connection in that country, and the price already includes it.
Which formats can I download?
JSON, CSV, Excel, XML and HTML from the dataset, or through the API.
Can I use it from Claude, ChatGPT or Cursor (MCP)?
Yes. Add the actor to the Apify MCP server and ask your assistant to check how ChatGPT answers a question. The "Copy to your AI assistant" block above has the call.
Is it legal to scrape ChatGPT answers?
The actor reads only what chatgpt.com shows to any logged out visitor and never logs in. Courts in the United States have found that accessing publicly available data is unlikely to breach the Computer Fraud and Abuse Act (hiQ Labs v. LinkedIn), but a website's terms and privacy laws still apply to what you do with the data. Check them for your use and take legal advice if you are unsure.
Integrations
Use Make, n8n, Zapier, Google Sheets, Airtable or a webhook to send each new answer where you need it. Schedules and webhooks are in the actor's Integrations tab.
More from The Mine Works
Marketing, SEO and reviews
- Facebook Ad Library Scraper
- Google Ads Transparency Scraper
- Similarweb Scraper
- Google News Scraper
- Trustpilot Reviews Scraper
- Google Trends Scraper
- Semrush Scraper
- App Reviews Scraper
- Trustpilot Business Search
- TripAdvisor Reviews Scraper
- Username Checker
- Backlink Building Agent
Social media and video
Leads and business directories
Real estate
Science, health and government data
Jobs and hiring
Company and business data
E-commerce and marketplaces
Food and local services
Developer and AI tools
More tools
Support
Found a prompt that returns nothing or a field that comes back empty? Open an issue in the Issues tab with the input and the run ID, and we will look at it. For a field you need or a new source, email dmineworks@gmail.com.
Ask ChatGPT your prompts and get the full answer, the cited sources and your brand's mentions back as one row per prompt, with no account and no API key.
ChatGPT and OpenAI are trademarks of their owner. This actor is not affiliated with, endorsed by or sponsored by OpenAI.

