Pricing

$9.00 / 1,000 pages

Go to Store

GPT Scraper

Try for free

Developed by

Jakub Drobník

Extract data from any website and feed it into GPT via the OpenAI API. Use ChatGPT to proofread content, analyze sentiment, summarize reviews, extract contact details, and much more.

4.4 (7)

Pricing

$9.00 / 1,000 pages

103

Total users

5.9K

Monthly users

150

Runs succeeded

>99%

Last modified

5 months ago

Lead generation

Back to issues Create new issue

Json schema is ignored and invalid output is generated

Closed

maroon_herb opened this issue

I have a json schema with: "arrayName": { "type": "array", "minItems": 1, "items": { "type": "object", "properties": {

But it still generates: "arrayName": [] I have also instructed GPT to: If it isn't possible to generate at least one record in arrayName answer with 'skip this page'

Jakub Drobník (drobnikj)

Hey Markus,

the reason is that if you require formatted output, the GPT cannot answer you with text. So the "skip this page" will not work here.

You need to process the output and filtered out the items with empty arrays.

Have a nice day!

maroon_herb

Hi Jakub,,

That explains the "skip this page" part, but still not whyI get an empty array when that isn't allowed in the schema. "minItems": 1 means that an empty array isn't valid.

Jakub Drobník (drobnikj)

You are right, currently, we do not validate the output against the schema, we just check if it is JSON. The schema is mainly made for defining the format of the output.

We can add validation, but we will provide this option in pay as you go version of the scraper https://apify.com/drobnikj/extended-gpt-scraper.

maroon_herb

That sounds like a good solution. In the meantime, add information that the schema isn't used for validation since "If true, the answer will be transformed into a structured format based on the schema in the jsonAnswer attribute." indicates that it's going to adhere to the schema. :)

Add comment

Extended GPT Scraper

drobnikj/extended-gpt-scraper

Extract data from any website and feed it into GPT via the OpenAI API. Use ChatGPT to proofread content, analyze sentiment, summarize reviews, extract contact details, and much more.

Jakub Drobník

1.4K

4.6

GPT Browser

anchor/gpt-browser

A GPT browser to use OpenAI prompt on any website. Put a list of URLs and a prompt, then the GPT agent will give you the answer you need. Fast, easy, and not limited with OpenAI ChatGPT restrictions. The best way to search and use GPT on large number of websites. Upload Excel or CSV. Screenshots 📸

Anchor

Auto GPT

lukaskrivka/auto-gpt

Run Auto GPT sessions directly on Apify. No OpenAI account or API token is required! Store parsed thoughts into datasets for later analysis.

Lukáš Křivka

198

GPT Search

tri_angle/gpt-search

Send queries to ChatGPT and retrieve structured answers with full source citations. Easily integrate into your tools or workflows for flexible, scalable AI-powered solutions.

Tri⟁angle

Pinecone GPT Chatbot

tri_angle/pinecone-gpt-chatbot

Pinecone GPT Chatbot combines OpenAI's GPT models with Pinecone's database to generate insightful responses. Its interactive chatbot interface presents precise and comprehensive answers to user queries. Benefit from semantic understanding, efficient workflows, and enriched knowledge integration!

Tri⟁angle

4.6

Universal AI GPT Scraper

louisdeconinck/ai-gpt-scraper

Transform any website into structured data with AI-powered extraction. This versatile tool combines advanced web scraping with intelligent content analysis to deliver clean, customized JSON output - perfect for automating data collection from any web source.

Louis Deconinck

5.0

🔍 GPT Search [Private API]

openapi/gpt-search-private-api

Use OpenAI's GPT4o Search mode via API! No cookie or proxy is required. Fast, cheap and reliable.

Open API

5.0

GPT-2 text generation

jirimoravcik/gpt2-text-generation

This actor uses the GPT-2 language model to generate text.

Jiří Moravčík

227

Convert GPT Image

jwkwjw/my-actor-7

This actor was build to help transpose Base 64 code (the output of Open AI's new imaging model). It accesses the data store of the API call to create a URL based result from Apify.

CMM

OpenRouter - Unified LLM Interface for ChatGPT, Claude, Gemini

xyzzy/open-router

Use the OpenRouter platform to choose the best and most cost effective model for your prompts utilizing a standardized interface (including ChatGPT, Claude, Gemini, Llama, Mistral, and more). See instructions for creating an OpenRouter account and API key.