/ai-web-scraping-scrapegraph

name: ai-web-scraping-scrapegraph
description: AI-powered web scraping - extract data using natural language prompts
source: orthogonal

ScrapeGraph AI - Intelligent Web Scraping

Setup

Read your credentials from ~/.gooseworks/credentials.json:

export GOOSEWORKS_API_KEY=$(python3 -c "import json;print(json.load(open('$HOME/.gooseworks/credentials.json'))['api_key'])")
export GOOSEWORKS_API_BASE=$(python3 -c "import json;print(json.load(open('$HOME/.gooseworks/credentials.json')).get('api_base','https://api.gooseworks.ai'))")

If ~/.gooseworks/credentials.json does not exist, tell the user to run: `npx gooseworks login`

All endpoints use Bearer auth: `-H "Authorization: Bearer $GOOSEWORKS_API_KEY"`

Extract web content using AI with natural language prompts.

Capabilities

• **Start SmartScraper**: Extract content from a webpage using AI by providing a natural language prompt and a URL
• **Start SearchScraper**: Start a new AI-powered web search request
• **Scrape**: Extract raw HTML content from web pages with JavaScript rendering support
• **Start SmartCrawler**: Start a new web crawl request with AI extraction or markdown conversion
• **Start Sitemap**: Extract all URLs from a website sitemap automatically
• **Start Markdownify**: Convert any webpage into clean, readable Markdown format
• **Get SearchScraper Status**: Get the status and results of a previous search request (free)
• **Get Markdownify Status**: Check the status and retrieve results of a Markdownify request (free)
• **Get Sitemap Status**: Check the status and retrieve results of a Sitemap request (free)
• **Get SmartCrawler Status**: Get the status and results of a previous smartcrawl request (free)
• **Get SmartScraper Status**: Check the status and retrieve results of a SmartScraper request (free)

Usage

Start SmartScraper

Extract content from a webpage using AI by providing a natural language prompt and a URL.

Parameters:

• user_prompt* (string) - Natural language description of what information you want to extract from the webpage.
• website_url* (string) - The URL of the webpage you want to extract information from. You must provide exactly one of: website_url, website_html, or website_markdown.
• website_html (string) - Raw HTML content to process directly (max 2MB). Mutually exclusive with website_url and website_markdown. Useful when you already have HTML content cached or want to process modified HTML.
• headers (object) - Optional custom HTTP headers to send with the request. Useful for setting User-Agent, cookies, authentication tokens, and other request metadata. Example: {"User-Agent": "Mozilla/5.0...", "Cookie": "session=abc123"}
• output_schema (object) - Optional schema to structure the output. If provided, the AI will attempt to format the results according to this schema.
• stealth (boolean) - Enable stealth mode to bypass bot protection using advanced anti-detection techniques. Adds +4 credits to the request cost
• website_markdown (string) - Raw Markdown content to process directly (max 2MB). Mutually exclusive with website_url and website_html. Perfect for extracting structured data from Markdown documentation, README files, or any content already in Markdown format.
• total_pages (number) - Optional parameter to enable pagination and scrape multiple pages. Specify the number of pages to extract data from. Default: 1 Range: 1-100
• number_of_scrolls (number) - Optional parameter for infinite scroll pages. Specify how many times to scroll down to load more content before extraction. Default: 0 Range: 0-50
• render_heavy_js (boolean) - Optional parameter to enable enhanced JavaScript rendering for heavy JS websites (React, Vue, Angular, SPAs). Use when standard rendering doesn’t capture all content. Default: false
• mock (boolean) - Optional parameter to enable mock mode. When set to true, the request will return mock data instead of performing an actual extraction. Useful for testing and development. Default: false
• cookies (object) - Optional cookies object for authentication and session management. Useful for accessing authenticated pages or maintaining session state. Example: {"session_id": "abc123", "auth_token": "xyz789"}
• steps (array) - Optional array of interaction steps to perform on the webpage before extraction. Each step is a string describing the action to take (e.g., “click on filter button”, “wait for results to load”). Example: ["click on search button", "type query in search box", "wait for results"]

curl -s -X POST $GOOSEWORKS_API_BASE/v1/proxy/orthogonal/run \
  -H "Authorization: Bearer $GOOSEWORKS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"api":"scrapegraph","path":"/v1/smartscraper"}'
  "website_url": "https://example.com/products",
  "user_prompt": "Extract all product names and prices"
}'

Start SearchScraper

Start a new AI-powered web search request

Parameters:

• user_prompt* (string) - The search query or question you want to ask. This should be a clear and specific prompt that will guide the AI in finding and extracting relevant information. Example: “What is the latest version of Python and what are its main features?”
• headers (object) - Optional headers to customize the search behavior. This can include user agent, cookies, or other HTTP headers. Example: { "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36", "Cookie": "cookie1=value1; cookie2=value2" }
• output_schema (object) - Optional schema to structure the output. If provided, the AI will attempt to format the results according to this schema. Example: { "properties": { "version": {"type": "string"}, "release_date": {"type": "string"}, "major_features": {"type": "array", "items": {"type": "string"}} }, "required": ["version", "release_date", "major_features"] }
• mock (string) - Optional parameter to enable mock mode. When set to true, the request will return mock data instead of perform