Pricing

from $8.00 / 1,000 extracted pages

Try for free

Go to Apify Store

AI Web Scraper - Extract Structured Data & URL to JSON

Try for free

AI scraper and AI extractor: list the fields or describe what you want, and an LLM reads each page and returns JSON. No selectors or code. One row per list item, follows detail links. $0.01/page.

Pricing

from $8.00 / 1,000 extracted pages

Rating

0.0

(0)

Developer

Yukai Lin

Actor stats

Bookmarked

Total users

Monthly active users

3 hours ago

Last modified

Categories

Automation

Developer tools

URLs or domains

urls

Optional

Paste a page URL here, then name the Fields you want or write Instructions. One URL per line; the book URL is a sample you can run now. Bare domains such as stripe.com also work. A line that is not a URL becomes a free error row; the other URLs continue. You can also use startUrls or a datasetId.

Type:array

URLs (request list, optional)

startUrls

Optional

Same as "URLs or domains", in Apify's request list format (also accepts a link to a text file with URLs). Kept for existing integrations; new integrations should use "urls".

Type:array

Ready-made fields

preset

Optional

Want to skip writing fields? Choose Property, Hotel or Menu, then clear the sample Fields below. Each preset supplies its columns and row mode; extra Fields are merged and your types win. Use pages you own or may access.

Type:string

Options:

propertyhotelmenu

Fields to extract

fields

Optional

Name what to extract: {"title":"string","price":"number","in_stock":"boolean"}. The sample matches the book URL above; edit both for your page. Missing values return null. Clear this sample when using a preset or when you want Instructions alone to choose fields. Types: string, number, integer, boolean, array.

Type:object

Instructions (or just a prompt)

instructions

Optional

Or write one sentence, such as "Get each product's name, price and rating". Clear the sample Fields first if you want the AI to choose columns from this sentence. With no Fields or JSON Schema, it chooses once from the first page, reuses the columns for this run and saves them as SCHEMA. You can also use this for hints such as "Prices in USD".

Type:string

Prompt (same as Instructions)

prompt

Optional

Same as "Instructions"; used only when Instructions is empty. Lets you paste the input JSON of Apify's AI Web Scraper (startUrls + prompt) and run it unchanged. Its crawl options (extractionMode scout/agentic, maxPagesToVisit, maxCrawlDepth) are ignored: only the start URLs are read, or turn on "Follow links to detail pages".

Type:string

Output

outputMode

Optional

"One row per page": one data object per URL. "One row per item": for listing pages (products, jobs, search results, team members), every item on the page becomes its own row with url and itemIndex. Still charged per page. Values: page = One row per page; items = One row per item on the page (lists).

Type:string

Default:page

Options:

pageitems

Flat columns (for Google Sheets, n8n, CSV)

flatOutput

Optional

Put every extracted field in its own column instead of inside "data". Nested objects become columns joined with _ (price.amount → price_amount), lists of values become one text cell ("red, blue"), lists of objects become a JSON text cell. Every row has the same columns in the same order, failed pages included (empty values, recordCount 0). recordIndex is the item's position on the page and the page title is in pageTitle. Same number of rows and same price as without it.

Type:boolean

Default:false

Follow links to detail pages

followLinks

Optional

Treat the start URLs as list pages (category, search results, index): read each one for free, then extract data from the pages it links to (e.g. every product page). The start page itself is not extracted. Each extracted page is charged as usual.

Type:boolean

Default:false

Link pattern

linkPattern

Optional

Which links to follow. A glob such as https://books.toscrape.com/catalogue/*/index.html (* = any text within one path segment, ** = across segments), or plain text the link must contain, such as /product/. Separate several with commas. Empty = every link to the same site (menus included, so a pattern is recommended).

Type:string

Max pages per start URL

maxPagesPerStartUrl

Optional

At most this many linked pages are extracted from each start page (in page order). Caps the cost.

Type:integer

Minimum:1

Maximum:1000

Default:20

JSON Schema (advanced)

jsonSchema

Optional

Optional full JSON Schema (type: object) for nested output. When set, it replaces "Fields". In item mode it describes one item.

Type:object

Dataset ID (optional)

datasetId

Optional

Read the URLs from another Actor run's dataset, e.g. a crawler or a search results scraper. Values that are not URLs are extracted from as text.

Type:string

Dataset field

datasetField

Optional

Field with the URLs, e.g. "url" or "website" (dot paths such as "company.url" work). Empty = the first URL field found.

Type:string

Page fetching (HTTP or browser)

mode

Optional

"Auto" reads pages over plain HTTP and switches to a real browser when a site blocks it or needs JavaScript. Values: auto = Auto (recommended); fast = Fast (HTTP only); browser = Browser (JavaScript rendering).

Type:string

Default:auto

Options:

autofastbrowser

Content selector (CSS)

cssSelector

Optional

Optional. Only read this part of each page, e.g. "main". Helps with very long pages.

Type:string

Parallel pages

maxConcurrency

Optional

How many pages are processed at the same time.

Type:integer

Minimum:1

Maximum:20

Default:5

Time limit per page (seconds)

itemTimeoutSecs

Optional

A page that takes longer (reading + AI) is stopped and returned as a free error row with errorType "timeout". In list mode, the items from the parts that finished in time are kept. 0 = no limit.

Type:integer

Minimum:0

Maximum:900

Default:180

AI model

aiModel

Optional

Same price for every model. Auto uses Llama 3.3 70B for one row per page and Qwen3 30B for list pages (one row per item), where it writes long lists about 3 times faster with the same accuracy in our tests.

Type:string

Default:auto

Options:

autollama-3.3-70bqwen3-30bllama-4-scout

Plain HTTP requests from

httpVia

Optional

Some sites block requests from data centers. Auto retries from a second network before falling back to a real browser. If both are refused, Auto also tries our second server (Oracle, different IP) before a real browser.

Type:string

Default:auto

Options:

autobackenddirectvm

Proxy (optional)

proxyConfiguration

Optional

Only used for requests sent from Apify's network. Apify proxy usage is billed to your Apify account.

Type:object

AI Web Scraper

apify/ai-web-scraper

AI-first web scraper that extracts structured data from any website using natural-language prompts. No programming knowledge required. No hard-coded logic that breaks when a website changes.

Apify

10K

4.9

(15)

AI Web Scraper - Powered by Crawl4AI

raizen/ai-web-scraper

A blazing-fast AI web scraper powered by Crawl4AI. Perfect for LLMs, AI agents, AI automation, model training, sentiment analysis, and content generation. Supports deep crawling, multiple extraction strategies and flexible output (Markdown/JSON). Seamlessly integrates with Make.com, n8n, and Zapier.

Raizen Technology

357

1.0

(1)

AI Web Scraper — Structured Data Extraction from Any Website

oneary/ai-powered-data-extractor

Extract structured data from any webpage using AI. Define your schema and the AI identifies relevant content — no selectors or coding needed. Handles products, reviews, contacts, and custom fields.

Luan M.

AI-Powered Smart Web Scraper

cloud9_ai/ai-web-scraper

Intelligent content extraction from any website using Crawlee + AI. Auto-detects structure, adapts to layout changes, handles JavaScript rendering. No custom code needed. Extract articles, products, listings from 1000s of pages.

cloud9

Web Content Extractor API — URL to JSON

george.the.developer/web-content-extractor-api

Extract structured JSON from any webpage. Articles, products, recipes, jobs. Auto-detects content type. Returns metadata, headings, images, links. For AI agents and RAG.

George Kioko

Smart AI Web Scraper

cockroachapi/smart-ai-web-scraper

Unlock the power of Smart AI Web Scraper! Efficiently scrape dynamic content, simulate browser behavior, and extract targeted data.

Cockroach API

5.0

(2)

AI Web Agent

apify/ai-web-agent

Use natural language prompts to browse the web, click on elements, fill and submit forms, extract data, and take screenshots using the OpenAI API.

Apify

1.9K

5.0

(9)

Scrape GPT - Universal AI Web Scraper Agent

paradox-analytics/scrape-gpt---universal-ai-web-scraper-agent

AI-powered universal web scraper that works on ANY website without configuration. Extract data from e-commerce, news sites, social media, and more using intelligent LLM-based field mapping. Features JSON-first extraction, automatic pagination, anti-bot bypass, and cost-effective caching.

Paradox Analytics

AI Web Scraper — Structured Data Extraction

muhammadafzal/ai-web-extractor

Extract structured JSON from public webpages using your own field schema. No CSS selectors. Ideal for products, jobs, articles, listings, RAG, and agents.

Muhammad Afzal

AI Web Data Extractor

datacortex/ai-web-data-extractor

Extract structured data from any website by saying what you want in plain English. Products, prices, jobs, leads, news, tables or direct answers, as clean JSON, CSV or Excel. No selectors, no code. Works with AI agents via MCP. Pay per page; failed pages are free.