# Changelog of PDF & Document to Markdown: OCR, Tables, RAG Chunks (`vimplement/document-to-markdown`) Actor

- **URL**: https://apify.com/vimplement/document-to-markdown/changelog.md
- **Full Actor documentation**: https://apify.com/vimplement/document-to-markdown.md

## Changelog

### 1.2 (2026-10-15)

- Upload files directly from the input form, with the same supported formats and charging as document URLs. Links and uploads are merged and deduplicated.
- Read private Apify uploads with the run token only after an anonymous request is refused; never forward it to other hosts.
- Prefill an 11-page public-domain NOAA report instead of a one-page placeholder. Prefill is not a default and is not added to upload-only API calls.
- Explain uploaded-file retention separately from temporary processing and saved output.

### 1.1 (2026-10-12)

- Output schema: links to the documents dataset and the run summary, required for the Apify Store.

### 1.0 (2026-10-12)

Initial release.

- PDF (text and scanned), images (PNG/JPG/TIFF, multi-page TIFF), DOCX, PPTX, XLSX, CSV, TXT/Markdown and HTML to Markdown or text.
- PDF layout analysis: reading order (multi-column), headings, lists, tables (Markdown + CSV).
- OCR with Tesseract only on pages without a text layer (or always), 12 languages preinstalled, others downloaded on demand; per-page confidence.
- RAG chunks with token sizes, overlap, page ranges and section headings.
- Pay per event: `document-extracted` (successes only) and `ocr-page`; stops cleanly at the user's maximum charge.
- Download guard: http(s) only, private and internal addresses refused on every redirect, size and time limits.