GPTs Scraper
The GPTs Scraper is an automated tool that allows extracting detailed information about any public GPT via its unique gizmoId. By simply inputting a gizmoId, the scraper will output a structured JSON file containing metadata about the target GPT. This provides an easy way to gather intel on ChatGPT extensions for analysis or research purposes.
Inputting a Gizmo ID
The GPTs Scraper accepts a single input – the unique gizmoId associated with a public GPT. This ID is contained in the GPT's URL on ChatGPT, after “/g/”.
For example, for the GPT at https://chat.openai.com/g/g-EgRT6abr4-gpt-explorer, which is a GPT explorer, that can help you find out the best GPT on ChatGPT and developed by GPTsApp.io the gizmoId would be:
g-EgRT6abr4
By inputting this ID into the scraper, it can locate the target GPT and extract its metadata.
Understanding the JSON Output
The scraper outputs a JSON file containing comprehensive information about the GPT. This includes:
- GPT name, description, author details
- Configuration like model used, instructions
- Metadata like creation date, usage stats
- Associated tools and services
- And more
Having all this data in a structured, machine-readable JSON format makes it easy to parse and analyze programmatically.
Sample JSON Output
Here is an example of the JSON output for a hypothetical GPT:
{
"gizmo": {
"id": "g-EgRT6abr4",
"organization_id": "org-eVJFbTXEADCyFwCzIkZJTaIx",
"short_url": "g-EgRT6abr4-gpt-explorer",
"author": {
"user_id": "user-I4jEsKxow4YaAGFePprBYOp2",
"display_name": "gptsapp.io",
"link_to": "https://gptsapp.io",
"selected_display": "website",
"is_verified": true
},
"voice": { "id": "ember"},
"workspace_id": null,
"model": null,
"instructions": null,
"settings": null,
"display": {
"name": "GPT Explorer",
"description": "A guide to recommending GPTs using user queries and API.",
"welcome_message": "Hello! I'm GPT Explorer. How can I assist you today?",
"prompt_starters": [
"Recommend a GPT for learning languages",
"What are the top GPTs for travel advice?",
"Find me a GPT for cooking recipes",
"Suggest a GPT for coding tutorials"
],
"profile_picture_url": "https://files.oaiusercontent.com/file-TQqtVSiCTBpzjkGw7HVLkaN0?se=2123-12-02T06%3A38%3A44Z&sp=r&sv=2021-08-06&sr=b&rscc=max-age%3D1209600%2C%20immutable&rscd=attachment%3B%20filename%3Da84cb259-3ef2-458b-b83c-b3a84a170c1f.png&sig=CW4WJ%2B/p9PaQqnzM1Z%2BavtFT/UHeCqAIsFD/ZyKVv1E%3D",
"categories": ["other"]
},
"share_recipient": "link",
"updated_at": "2023-12-26T07:34:11.785757+00:00",
"last_interacted_at": null,
"tags": ["public","reportable","uses_function_calls"],
"version": null,
"live_version": null,
"training_disabled": null,
"allowed_sharing_recipients": null,
"review_info": null,
"appeal_info": null,
"vanity_metrics": {
"num_conversations_str": "5",
"num_pins": 0,
"num_users_interacted_with": 0
}
},
"tools": [
{
"id": "gzm_cnf_4FV9fcScgNMw47GIT4cGrdYX~gzm_tool_oOnVBMhnOJ82SL1HhI88uklI",
"type": "plugins_prototype",
"settings": null,
"metadata": {
"action_id": "g-770ffd1996213259c666756e704fc8263cd87b56",
"domain": "api.smartproxt.com",
// 其他 metadata
}
},
{
"id": "gzm_cnf_7FeZ1BrGP1vgewZgGSCLu7eH~gzm_tool_xsDPbTo7C3IhbWyRlzwcd7El",
"type": "python",
"settings": null,
"metadata": null
},
{
"id": "gzm_cnf_VRQOeDHb85H6CGV1v9NNniQM~gzm_tool_uqfnomcmSwddUOrq0SLxWKrS",
"type": "browser",
"settings": null,
"metadata": null
},
{
"id": "gzm_cnf_VRQOeDHb85H6CGV1v9NNniQM~gzm_tool_cZwFEan0DJZ5K2VO00HOJT3Y",
"type": "dalle",
"settings": null,
"metadata": null
}
],
"files": [],
"product_features": {
"attachments": {
"type": "retrieval",
"accepted_mime_types": [
"application/msword","application/vnd.openxmlformats-officedocument.presentationml.presentation","text/x-script.python","text/x-c","application/x-latext","text/x-php","text/markdown","application/json","text/x-csharp","application/vnd.openxmlformats-officedocument.wordprocessingml.document","text/html","text/x-java","text/x-sh","text/x-typescript","text/x-c++","application/pdf","text/plain","text/javascript","text/x-tex","text/x-ruby"
],
"image_mime_types": [
"image/gif","image/png","image/jpeg","image/webp"
],
"can_accept_all_mime_types": true
}
}
}
This covers all the key details exposed by the scraper ranging from authorship information to tool integrations.
Information Contained in the JSON File
The main sections of information included in the JSON output are:
gizmo – Metadata about the GPT itself like name, description, author, stats etc.
tools – Any tools associated with the GPT like plugins, Python scripts etc.
files – References to files uploaded for the GPT.
product_features – Configuration of features like attachments.
This comprehensive set of details enables thoroughly analyzing and understanding a GPT's purpose, creation, configuration and usage based on the JSON output.
Getting Started with the Scraper
To start using the GPTs Scraper:
- Obtain the gizmoId of the target GPT, contained in its ChatGPT URL
- Input this ID when prompted by the scraper
- The scraper will fetch the GPT's metadata and output it as a JSON file
- The JSON contains a wealth of information about the GPT as covered in this guide
Using the scraper requires no coding or technical expertise. Simply pass a valid gizmoId and receive structured GPT intel via JSON.
Use Cases
The GPTs Scraper enables several useful applications:
- GPT analysis – Study GPT configurations, usage patterns, authorship trends
- GPT archiving – Create dataset of GPT metadata for research
- Competitive intel – Analyze competitors' GPT strategies and offerings
- GPT discovery – Search for GPTs like GPT store based on metadata like tags, author etc
- Sentiment analysis – Extract unstructured text from GPTs for text analysis
These are just some examples of how the structured JSON output can inform both analysis of broader GPT trends as well as individual GPT performance and characteristics.
Limitations
While powerful, the GPTs Scraper does have some key limitations to note:
- Only works for public GPTs – private or restricted GPTs cannot be scraped
- Requires knowing the exact gizmoId, not just name or URL
- Only extracts data exposed by ChatGPT's API
- Does not capture conversation logs or usage data
- May fail if ChatGPT API changes
Despite these constraints, the scraper provides extremely useful metadata for the majority of public GPTs via an easy-to-ingest JSON format.
FAQs
Does the scraper work for private GPTs?
No, it can only access information about public GPTs available on ChatGPT.
What if I don't know the exact gizmoId?
You need the full unique gizmoId contained in the GPT's URL in order for the scraper to locate it and extract metadata.
What format is the output in?
The scraper returns a JSON file containing the GPT metadata. This allows programmatic processing and analysis of the data.
What can I analyze with the scraped GPT data?
Many aspects like author trends, tool usage, configurations, tags, temporal analysis and more. The structured JSON enables flexibly slicing and dicing the data.
Can I scrape conversation logs too?
No, the scraper only returns metadata exposed through ChatGPT's API, not actual conversations or usage logs.
This covers the key capabilities and value proposition of the GPTs Scraper for easily extracting a wealth of knowledge about public ChatGPT extensions. With some clever effort, the JSON output can unlock even deeper GPT insights.
