Technical previewShare feedback

Search

Search documentation

0.1.0

Published Oct 30, 2026

Features

  • add html.externalLinks and attribute lists for md.headingId (473b5ec)
  • add md.headingId, md.strip, html.unwrapImages, html.extractToc, code rules in md.component, file data and ANSI highlighting (f0b05ae)
  • add the core processor pipeline with unified's use, data and freeze (4a65027)
  • align the bindings' host plugin API and hold it to shared conformance cases (4b8f1ae)
  • bootstrap the arena tree, the processor and Markdown to HTML (45afd1e)
  • convert HTML to Markdown and add GFM (b708ce7)
  • expose the processor in the Node and Python bindings (400a7ab)
  • give JavaScript and Python plugins vfile's file and fail runs on a returned error (13a95f4)
  • html: add html.highlight on giallo (36501e7)
  • html: add html.katex, rendering math as rehype-katex does, natively on katex-rs (ba1264f)
  • html: add html.raw and html.format (33700b3)
  • html: parse HTML as parse5 and hast-util-from-html do (f0afe97)
  • html: run Shiki's notation, meta-highlight and line-number transformers, inline code and color replacements natively in html.highlight (3716429)
  • html: sanitize and slug hast (eb083fb)
  • html: serialize every hast-util-to-html option (b6c4083)
  • let each thread keep up to 256 MB between documents, configurable with setRetainedMemory (5f0e90c)
  • let JavaScript and Python plugins join a pipeline as transformers (03244b1)
  • make each Rust plugin a struct of its options, replacing the constructor functions (fcf4eb3)
  • md: add math, directive, frontmatter and definition list extensions (739178b)
  • md: add md.component, turning directives, alerts and nodes into MDX components by rules (3aeebcd)
  • md: add md.toc, md.packageCommand and frontmatter data, with github-slugger in core (0581318)
  • md: apply unist data and passThrough in md.toHtml and fix parser gaps (3a58dc6)
  • md: parse block quotes and lists (2ea6249)
  • md: parse character escapes, references, emphasis and hard breaks (912462e)
  • md: parse code spans, fenced code and indented code (3770324)
  • md: parse setext headings, links, definitions, autolinks and raw HTML (28abcfe)
  • md: serialize mdast to Markdown and replace gemoji shortcodes (51d7831)
  • mdx: add md.mdx and mdx.compile (e5ad62d)
  • prose: add prose parsing, retext plugins and the md and html bridges (8ae05e0)
  • python: make process and run synchronous, with aprocess and arun as their coroutines (2a6a963)
  • rename the prose domain to nl, with md.toNaturalLanguage and html.toNaturalLanguage (95042ba)

Bug Fixes

  • html: export ToNaturalLanguage without the highlight feature (d9d9c84)
  • md: match micromark on tabs, laziness, list ends and code spans (4adf21f)
  • mdx: add the mdx.compile plugin an ignore rule kept out (1726e83)

Performance Improvements

  • md: build trees at exact size and scan escapes with memchr (23e3f93)
  • md: halve memory and speed up parsing 1.25–2× (a87fa5b)
  • md: parse Markdown in one pass, MDX included (1d05eee)
  • md: skip plain text and texts without literals under GFM (fd73efe)
  • md: skip plain text in one run and find line ends with memchr (d72f312)
  • md: speed up parsing and HTML output 2–3× (6a30a3f)

Created with ❤ and Livemark