# Changelog of PDF & Document to Markdown — RAG Chunks, Tables, OCR (`frayess_mosbehi/doc-to-markdown-rag`) Actor

- **URL**: https://apify.com/frayess_mosbehi/doc-to-markdown-rag/changelog.md
- **Full Actor documentation**: https://apify.com/frayess_mosbehi/doc-to-markdown-rag.md

## Changelog

### 0.2

- New input `layout` (`fast` default | `accurate`). Fast layout converts PDFs without ML models: text-layer pages via
  pdfplumber (font-size headings, ruled tables, two-column reading order), scanned pages via Tesseract directly.
  Mixed PDFs are handled page by page. `accurate` uses Docling's AI layout model for every page and is billed at the OCR page price.
- Output schema added (required for publishing to the Store).
- Docling conversion pinned to 1 thread to match Apify's 1 vCPU per 4 GB.

### 0.1 — unreleased

- First version: PDF, DOCX, PPTX, XLSX, HTML → Markdown / JSON / text, tables, RAG chunks.
- Automatic OCR (Tesseract) only for pages without a text layer.
- Pay per page: `page-converted`, `ocr-page-converted`; no start or per-file fee.
- SSRF protection, size cap, magic-byte detection, zip-bomb guard, per-document timeout.