# Changelog of Website & URL to Markdown Converter for LLMs (`cleandata-labs/url-to-markdown`) Actor

- **URL**: https://apify.com/cleandata-labs/url-to-markdown/changelog.md
- **Full Actor documentation**: https://apify.com/cleandata-labs/url-to-markdown.md

## Changelog

All notable changes to URL to Markdown Converter. Dates are UTC. Small internal refactors and documentation-only updates are omitted.

### 0.1.11 (2026-10-28)

#### Fixes

- Page titles and descriptions keep literal `&` sequences exactly as published (no double HTML-entity decoding in metadata).
- Plain-text output keeps underscores and asterisks that belong to words or code (`snake_case`, `*args`, `2*3`) and keeps code blocks verbatim.
- Optional inputs sent as `null`, empty strings or invalid values fall back to their defaults instead of failing the run.

### 0.1.10 (2026-10-28)

#### Fixes

- Links and text containing `&` are no longer mangled by HTML-entity decoding (for example `&section=` in URLs).

### 0.1.9 (2026-10-28)

#### Fixes

- RAG chunks are split on word boundaries and no chunk consists only of overlap text.
- In-page reference links (for example Wikipedia `#cite_note` links) keep the full page URL, so they stay valid outside the page.

### 0.1 (2026-10-27)

#### Initial release

- Convert public web pages and PDFs to clean Markdown, plain text or HTML with title, author, date, language, description and canonical URL.
- Main-content or full-page extraction, CSS selectors to keep or remove elements, links, tables and optional images.
- Small crawls by link depth or sitemap with include/exclude patterns; duplicate pages skipped.
- Headless-browser fallback only when a page needs JavaScript; token-sized chunks with overlap for RAG.
- Respects robots.txt; blocked, 4xx, robots-disallowed and duplicate URLs are free.
