Skip to main content
ScrapeGraphAI provides a collection of specialized graphs for different web scraping scenarios. Each graph is designed for specific use cases and offers unique capabilities.

Available Graphs

SmartScraperGraph

Extract structured data from single web pages using AI

SearchGraph

Search the internet and scrape multiple results

SpeechGraph

Generate audio summaries from scraped content

ScriptCreatorGraph

Generate custom scraping scripts with BeautifulSoup

DocumentScraperGraph

Extract information from text documents and markdown

SmartScraperMultiGraph

Scrape multiple URLs and merge results intelligently

DepthSearchGraph

Crawl websites with configurable depth and link following

OmniScraperGraph

Scrape and analyze both text and images from web pages

Choosing the Right Graph

Single Page Scraping

  • SmartScraperGraph: Best for extracting structured data from a single URL
  • OmniScraperGraph: When you need to analyze images along with text
  • DocumentScraperGraph: For processing plain text or markdown content

Multiple Pages

  • SmartScraperMultiGraph: Scrape multiple known URLs with merged results
  • SearchGraph: Search the internet first, then scrape results
  • DepthSearchGraph: Crawl an entire website following internal links

Special Purposes

  • SpeechGraph: Generate audio summaries from scraped content
  • ScriptCreatorGraph: Generate reusable Python scraping scripts

Comparison Table

GraphInputOutputBest ForAPI Key Required
SmartScraperGraphSingle URL/HTMLStructured dataGeneral web scrapingYes
SearchGraphSearch queryMerged resultsResearch and data aggregationYes
SpeechGraphSingle URLText + Audio fileContent summarizationYes (OpenAI)
ScriptCreatorGraphSingle URLPython scriptCode generationYes
DocumentScraperGraphText/MarkdownStructured dataDocument processingYes
SmartScraperMultiGraphMultiple URLsMerged dataMulti-source scrapingYes
DepthSearchGraphSingle URLDeep scraped dataSite-wide scrapingYes
OmniScraperGraphSingle URLText + Image descriptionsVisual content analysisYes (OpenAI GPT-4)

Common Configuration Options

All graphs share common configuration parameters:
graph_config = {
    "llm": {
        "model": "openai/gpt-4o-mini",  # or "ollama/llama3.2"
        "api_key": "your-api-key",      # for OpenAI models
        "temperature": 0,                # 0-1, lower = more deterministic
    },
    "verbose": True,                     # Enable detailed logging
    "headless": False,                   # Browser visibility
}

LLM Provider Support

OpenAI

"llm": {
    "model": "openai/gpt-4o-mini",
    "api_key": os.getenv("OPENAI_API_KEY"),
}

Ollama (Local)

"llm": {
    "model": "ollama/llama3.2",
    "temperature": 0,
    "base_url": "http://localhost:11434",
}
Most graphs support both OpenAI and Ollama models. However, SpeechGraph and OmniScraperGraph require OpenAI’s GPT-4 vision and TTS models.

Next Steps

SmartScraperGraph

Start with the most popular graph

Installation

Set up ScrapeGraphAI

Build docs developers (and LLMs) love