ChatGPT Marketplace Scraper

🔎 ChatGPT marketplace scraper for public GPT data

ChatGPT Marketplace Scraper collects public custom GPT listings from ChatGPT. Use this chatgpt marketplace scraper to search by phrase, browse categories or marketplace pages, or enrich direct GPT listing URLs without opening each listing by hand.

Each saved row is one public custom GPT listing. Rows can include the GPT ID, name, description, image URL, creator details, verified creator status, categories, tags, capabilities, prompt examples, usage metrics, review facts when exposed, listing status, timestamps, and discovery context.

You do not need a ChatGPT account, OpenAI API key, cookies, private token, or browser extension. For a quick first run, keep the prefilled data analysis target, leave Discovery order on source relevance, and use a small Maximum GPTs per discovery target value before raising the limit.

✅ What this Actor does

• Searches public custom GPT listings from search phrases such as data analysis, invoice assistant, or marketing.
• Scrapes direct public GPT listing URLs such as https://chatgpt.com/g/g-2fkFE8rbu-dall-e.
• Reads public marketplace, category, or tag targets when the source exposes them.
• Saves one dataset row per accepted public custom GPT listing.
• Deduplicates the same GPT inside a run by its stable public GPT ID.
• Keeps source-native labels and metrics instead of inventing categories, scores, summaries, or recommendations.
• Skips private, unavailable, failed, or unsupported targets instead of mixing error rows into the dataset.

This Actor is built for GPT marketplace research, AI tool directories, creator research, category monitoring, competitor lists, and repeatable custom GPT exports. It focuses on public listing facts only. It does not scrape private chats, private GPT configuration, builder dashboards, account analytics, or logged-in-only workspace data.

📦 Data you can extract

Each dataset item can include:

• gptId and name
• description and imageUrl
• creator.name, creator.url, creator.verified, and creator.organization
• categories, tags, and capabilities
• model, voice, and promptExamples
• conversationCount, rating, reviewCount, and reviewSummary
• listingStatus, publishedAt, and updatedAt
• discovery.targets, discovery.sourceUrl, and discovery.rank

Some fields can be null or empty when ChatGPT does not expose that fact for a public listing. Direct listing URL targets return discovery: null because the submitted listing is already the source.

🎯 Target options

Add one target per row in GPT marketplace targets.

Supported target formats:

• Direct GPT listing URL: https://chatgpt.com/g/g-2fkFE8rbu-dall-e
• Search phrase: data analysis
• Category or tag name: productivity
• Marketplace page URL: https://chatgpt.com/gpts

Maximum GPTs per discovery target caps rows saved from search, category, tag, and marketplace page targets. Direct listing URLs return one matching GPT.

Discovery order chooses the source order for discovery targets when available:

• Source relevance keeps the source's default result order.
• Popular first sorts discovered listings by source-native conversation count when available.
• Newest first uses source-supported date order for search and source-native update dates for other discovery rows.

🚀 How to run it

1. Open the Actor input.
2. Add GPT URLs, marketplace URLs, search phrases, category names, or tag names.
3. Keep Maximum GPTs per discovery target small for your first run.
4. Choose a discovery order only if you need a specific result order.
5. Run the Actor and open the dataset.

You can export the dataset as JSON, CSV, Excel, XML, RSS, or HTML. You can also run the Actor through the Apify API, schedule repeat runs, or connect results to webhooks and integrations.

🧾 Output example

{
	"gptId": "g-2fkFE8rbu",
	"name": "DALL-E",
	"description": "Create images from text prompts.",
	"imageUrl": "https://example.com/gpt-icon.png",
	"creator": {
		"name": "OpenAI",
		"url": "https://chatgpt.com/gpts/mine/openai",
		"verified": true,
		"organization": "OpenAI"
	},
	"categories": ["Productivity"],
	"tags": ["image generation", "design"],
	"capabilities": ["DALL-E", "Image generation"],
	"model": "GPT-4o",
	"voice": null,
	"promptExamples": ["Create a watercolor fox reading a book"],
	"conversationCount": 2500000,
	"rating": 4.7,
	"reviewCount": 1842,
	"reviewSummary": "Highly rated by users",
	"listingStatus": "active",
	"publishedAt": "2024-01-10",
	"updatedAt": "2026-06-01",
	"discovery": {
		"targets": ["data analysis"],
		"sourceUrl": "https://cse.google.com/cse?cx=b0b573346627a4812#gsc.tab=0&gsc.q=data+analysis&gsc.page=1&gsc.sort=date",
		"rank": 3
	}
}

💳 Pricing

This Actor uses pay-per-event pricing. You are charged once for each accepted public custom GPT listing saved to the dataset.

Failed targets, private targets, unavailable listings, unsupported inputs, skipped duplicates, and zero-row outcomes are not saved as result rows and are not charged as custom GPT listing events.

⚠️ Limits and caveats

• The Actor returns public marketplace facts that ChatGPT exposes at run time.
• Search phrase discovery uses public search discovery, then enriches accepted rows from ChatGPT public listing details.
• Source-native metrics can be missing, delayed, rounded, or unavailable for some listings.
• The Actor does not infer popularity, category, sentiment, recommendation quality, or review meaning.
• The Actor does not use your private ChatGPT or OpenAI account state.

❓ FAQ

🔐 Do I need a ChatGPT login?

No. The Actor is designed for public custom GPT listings and does not ask for ChatGPT cookies, an OpenAI API key, or a private token.

🎯 Can I scrape one known GPT?

Yes. Paste the public GPT listing URL in GPT marketplace targets. Direct listing targets return one matching row when the listing is public and available.

🔎 Can I search for GPTs by keyword?

Yes. Add a plain search phrase such as data analysis, sales assistant, or seo. The Actor discovers matching public custom GPT listings and then saves source-backed listing details.

📭 Why are some fields null?

ChatGPT does not expose every field for every public listing. The Actor keeps unavailable facts as null or empty arrays instead of guessing.

📝 Changelog

• 0.0: Initial release.

🆘 Support

For issues, questions, or feature requests, file a ticket and I'll fix or implement it in less than 24h 🫡

🔗 Other actors

• Chrome Extensions Scraper ↗ - Export Chrome Web Store extension listings for app and tool research.
• Product Hunt Scraper ↗ - Collect startup launch data, rankings, reviews, and website enrichment.
• RapidAPI Scraper ↗ - Search RapidAPI marketplace listings and enrich public API profiles.
• G2 Products Scraper ↗ - Collect public software product, rating, review, and vendor data from G2.
• Apple App Store Scraper ↗ - Export public iOS app metadata, charts, ratings, and reviews.

Made with ❤️ by Maxime Dupré