PDF to Markdown RAG-Ready Scraper

🚀 Convert complex PDF documents into clean, structured Markdown — perfectly optimized for RAG pipelines, LLM fine-tuning, and AI agents.

Why This Actor?

Extracting text from PDFs is easy, but extracting meaning is hard. This Actor is specifically tuned for the needs of modern AI:

Feature	Standard PDF Parsers	This Actor
Table Preservation	❌ Scrambled text	✅ Structured Markdown tables
Hierarchical Headings	❌ Flat text	✅ Nested sections (H1-H6)
Semantic Chunking	❌ Arbitrary splits	✅ Context-aware RAG chunks
Metadata Extraction	❌ Minimal	✅ Author, Title, Creator, Dates
RAG-Ready Output	❌ Full file only	✅ Chunked JSON for Vector DBs

🎯 RAG-Ready Output

Every PDF is broken down into semantically coherent chunks, ready to be indexed into Chroma, Pinecone, or Weaviate:

{
    "url": "https://example.com/report.pdf",
    "chunk": "### 3.1 Quarterly Results\nOur revenue grew by 15%...",
    "headings": ["3. Financial Growth", "3.1 Quarterly Results"],
    "docMetadata": {
        "title": "Annual Report 2024",
        "author": "Corporate Strategy Team",
        "pageCount": 42
    }
}

Key Features

• Structural Integrity: Preserves bold text, lists, and hierarchical structure.
• Premium OCR: Handles scanned PDFs and image-heavy documents (optional).
• Embedded Tables: Converts complex PDF tables into clean Markdown format.
• Smart Metadata: Automatically extracts document info for better context in RAG.
• Pay-Per-Event: No fixed monthly costs. You pay only for what you process.

🔗 LangChain Integration (Python)

from langchain.document_loaders import ApifyDatasetLoader
from langchain.docstore.document import Document

loader = ApifyDatasetLoader(
    dataset_id="YOUR_DATASET_ID",
    dataset_mapping_function=lambda item: Document(
        page_content=item["chunk"],
        metadata={
            "source": item["url"],
            "headings": " > ".join(item["headings"]),
            **item["docMetadata"]
        }
    ),
)
docs = loader.load()

Input Parameters

Field	Type	Description
urls	Array	List of PDF URLs to process
chunkSize	Number	Maximum characters per semantic chunk (default: 1000)
enableChunking	Boolean	Whether to split document into RAG chunks
includeMetadata	Boolean	Include original PDF metadata in output

Pricing

Pay per Event:

• Actor Start: $0.01 per GB of memory
• RAG-Ready Chunk: $0.001 per extracted chunk

Author

Built with ❤️ by HEDELKA for the AI Engineering community.

Questions? Open a GitHub issue or contact us on the Apify platform.