AST Website Content Crawler for RAG avatar

AST Website Content Crawler for RAG

Pricing

$3.00 / 1,000 results

Go to Apify Store
AST Website Content Crawler for RAG

AST Website Content Crawler for RAG

Advanced web crawler extracting token-chunked Markdown for LLMs & RAG. Features dual-engine hybrid parsing (IBM Docling & Trafilatura) to preserve tables and layouts. Delivers clean text split by native tiktoken counts with rich metadata (headings, levels, tokens) for instant vector search.

Pricing

$3.00 / 1,000 results

Rating

5.0

(2)

Developer

lukas schmeck

lukas schmeck

Maintained by Community

Actor stats

3

Bookmarked

3

Total users

0

Monthly active users

14 days ago

Last modified

Share

AI Web-to-Markdown Crawler

🚀 AI Web-to-Markdown Crawler

Production-Ready Web Extraction Pipeline for LLMs, RAG Systems & AI Knowledge Bases

Turn any website into clean, structured, AI-ready Markdown — optimized for embeddings, vector databases and Retrieval-Augmented Generation.

Python Apify RAG LLM

Why this Actor?

Most web crawlers collect HTML.

But AI systems don't need HTML.

They need clean, structured, context-preserving knowledge.

This Actor transforms websites into high-quality Markdown documents designed specifically for:

  • 🤖 LLM applications
  • 🔎 RAG pipelines
  • 🧠 AI agents
  • 📚 Knowledge bases
  • 🗄 Vector databases
  • 📖 Documentation search

Instead of feeding your AI:

❌ cookie banners
❌ navigation menus
❌ duplicated content
❌ broken HTML
❌ irrelevant website noise

You get:

✅ clean Markdown
✅ preserved structure
✅ meaningful chunks
✅ metadata-rich documents
✅ embedding-ready output


✨ Features

🧠 AI-Optimized Markdown Extraction

Uses a hybrid parsing system combining:

  • Trafilatura for fast high-quality article extraction
  • Docling for complex documents and structured content

The Actor automatically evaluates extraction quality and selects the best result.

⚡ Adaptive Parsing Intelligence

The crawler learns which parser performs best for each domain.

Example:

example.com → Trafilatura
documentation-site.com → Docling

Future crawls reuse successful strategies automatically.


🔥 Built for RAG Pipelines

Every page is converted into semantic chunks optimized for retrieval.

Each chunk includes:

{
"chunk_id": "...",
"text": "...",
"metadata": {
"title": "...",
"url": "...",
"heading_path": [],
"language": "...",
"tokens": 512,
"quality_score": 94.2
}
}

Perfect for:

  • OpenAI embeddings
  • Pinecone
  • Weaviate
  • Qdrant
  • ChromaDB
  • LangChain
  • LlamaIndex

🧩 Intelligent Chunking

Unlike simple text splitters, this Actor understands document structure.

Features:

✅ heading-aware splitting
✅ token-based limits
✅ table preservation
✅ code block protection
✅ context overlap
✅ deterministic chunk IDs

Your AI receives meaningful knowledge units instead of random text fragments.


🧹 Advanced Content Cleaning

Automatically removes:

  • cookie dialogs
  • navigation menus
  • advertisements
  • popups
  • scripts
  • irrelevant UI elements

Includes quality filtering to reject:

  • empty pages
  • spam pages
  • menu-only content
  • duplicate documents

🔄 Incremental Crawling

Large documentation websites can be crawled efficiently.

The Actor remembers:

  • previously processed content
  • failed URLs
  • successful parser strategies

Only changed content needs to be processed again.


🔐 Production Security

Built with enterprise-style protections:

✅ SSRF protection
✅ private network blocking
✅ cloud metadata protection
✅ URL normalization
✅ content-size limits
✅ retry handling

Safe for automated crawling workflows.


📊 Perfect Use Cases

AI Chatbots

Create a chatbot trained on:

  • documentation
  • help centers
  • company websites

Internal Knowledge Bases

Convert:

  • company documentation
  • manuals
  • wikis
  • technical pages

into searchable AI knowledge.

Market Research

Extract structured information from:

  • competitors
  • product pages
  • industry websites

Developer Documentation

Create datasets from:

  • API docs
  • GitHub documentation
  • technical references

🚀 Example Workflow

Website
|
AI Web-to-Markdown Crawler
|
Clean Markdown
|
Semantic Chunks
|
Embeddings
|
Vector Database
|
AI Assistant / RAG Application

Input Example

{
"startUrls": [
{
"url": "https://docs.example.com"
}
],
"maxPages": 50,
"maxDepth": 3,
"parserMode": "auto"
}

Output Example

{
"text": "# Authentication API\n\nThe API allows...",
"metadata": {
"source_type": "web_documentation",
"language": "en",
"word_count": 1240,
"reading_time_minutes": 6
}
}

Configuration

OptionDescription
maxPagesMaximum pages to crawl
maxDepthCrawl depth
parserModeauto / docling / trafilatura
chunkSizeMaximum tokens per chunk
chunkOverlapContext overlap
respectRobotsTxtRespect robots.txt
incrementalCrawlEnable delta crawling

Why not use a normal HTML scraper?

Traditional crawlers optimize for:

Collect everything

This Actor optimizes for:

Give AI the right information

The difference matters.

Bad input creates bad AI answers.

Better retrieval starts with better data.


Performance Highlights

Designed for real-world production workloads:

  • adaptive extraction
  • memory-safe document processing
  • duplicate detection
  • intelligent caching
  • scalable crawling architecture

Ideal For

⭐ AI developers
⭐ RAG engineers
⭐ SaaS builders
⭐ Data engineers
⭐ Automation specialists
⭐ Companies building AI assistants


Try it now 🚀

Turn your website into AI-ready knowledge in minutes.

No custom crawler required.

Just provide a URL and receive structured Markdown optimized for your next AI workflow.