Markdown parser, done right. Commonmark support, extensions, syntax plugins, high speed - all in one. Gulp and metalsmith plugins available. Used by Facebook, Docusaurus and many others! Use https://github.com/breakdance/breakdance for HTML-to-markdown conversion. Use https://github.com/jonschlinkert/markdown-toc to generate a table of contents.
Remarkable is a decade-old project (created 2014) that reached a stable v2.0.0 release in August 2019 — now nearly seven years without a formal release. The package.json shows an unpublished v2.0.1 bump, suggesting maintenance has stalled before release. The Travis CI badge (not GitHub Actions) and a devDependency list pinned to 2019-era tool versions (mocha ^6, rollup ^1, eslint ^6) reinforce that the project is in slow-life mode. With 116 open issues and no triage evidence, the issue queue is accumulating. Still functional and not archived, but squarely in maintenance-only territory.
Exceptional documentation for a parser library. The README covers installation (Node + CDN), ESM/CJS usage, all configuration options with defaults, three presets, syntax highlighting integration (hljs example), all syntax extensions (enabled and disabled), rule management API, typographer replacements, plugin system, and linkify. A live demo is available. Beyond the README, a dedicated docs/ directory holds six targeted documents: parser architecture, block parsing, core parsing, inline parsing, plugin authoring, and renderer. This is the kind of documentation that lets a developer be productive in under 30 minutes.
Clean separation of concerns: core parsing pipeline with block/inline/core rule chains, a renderer abstraction, and a plugin hook system. Rollup produces both ESM and CJS bundles with browser variants — a deliberate dual-distribution strategy. Runtime dependencies are minimal: argparse (CLI) and autolinker (linkify). Test suite uses mocha with nyc coverage and a CommonMark spec conformance runner (specsplit.js). The engines field requires Node ≥ 6.0.0, which is technically end-of-life ancient. CI tooling is Travis CI rather than GitHub Actions, and devDependencies are several major versions behind current — functional but showing age.
Last commit May 2024 (~2 years ago) with no release since v2.0.0 in 2019 is a red flag. The gap between the package.json version (2.0.1) and the published release (2.0.0) suggests maintenance commits are happening but not being shipped. 116 open issues with no visible triage cadence. PR merge cadence is unclear but the combination of stale tooling, no releases, and accumulating issues signals the project is in long-term passive maintenance. It is notable that remarkable was the direct inspiration for markdown-it, which has since become the ecosystem successor and absorbed most active development energy.
5,831 stars and 376 forks reflect genuine historical traction. Named downstream users include Facebook and Docusaurus (a Meta project with massive distribution). Available on both jsDelivr CDN and cdnjs, indicating broad browser-side use. A community plugin ecosystem exists on npm under the remarkable keyword. However, Docusaurus migrated from remarkable to mdx-js/remark in v2, so the named endorsement is partially historical. The star trajectory has likely plateaued given the ecosystem migration to markdown-it and unified/remark.
Overall: 3.4/5
Category: Markdown Parsing Engine
Known alternatives in vault: None — no markdown parsing or text processing repos exist in the vault.
Differentiation: Remarkable offers a fully configurable rule-chain architecture with named presets (commonmark, full), first-class typographer support, and a clean plugin API. It was the originator of the parse-then-render pipeline that markdown-it later refined. Compared to marked (simpler, less extensible) or unified/remark (more powerful but far heavier), remarkable occupies a middle ground. Its key weakness relative to current alternatives: markdown-it (its direct spiritual successor, maintained by the same Vitaly Puzrin who contributed here) is faster, more spec-compliant, and actively maintained — remarkable has no meaningful technical advantage over it today.
Gap or crowd: Gap — this would be the only markdown processing repo in the vault. However, the category itself has limited PAI strategic value unless the vault specifically processes markdown content at ingestion time.
Score: 2/5
Harvestable: The rule-chain plugin architecture (block/inline/core ruler pattern) is a well-executed pattern worth studying for any PAI pipeline that needs extensible text transformation. The typographer replacement table and GFM extension toggles are concrete examples of configurable text normalization. The benchmark harness comparing against marked and commonmark-reference is a useful testing pattern for parser work.
Integration path: Could serve as a markdown-to-HTML rendering step in a knowledge vault ingestion pipeline — converting .md notes to structured HTML for downstream extraction or indexing. CLI mode (cat myfile.md | remarkable) makes it trivially scriptable. However, markdown-it would be the better choice for new integration given active maintenance.
Overlap with existing: No existing vault repos serve markdown parsing. No overlap detected.
Adoption cost: Trivial for basic usage (npm install remarkable, two lines of JS). Moderate if building custom syntax rules or plugins. Significant if expecting long-term maintenance alignment — at that point, migrating to markdown-it would be the recommended path instead.
Remarkable is a historically important, well-documented, and once-dominant Markdown parser that has effectively been superseded in the ecosystem by its own successor, markdown-it. For a PAI vault, the repo has archival and educational value — particularly the rule-chain architecture and plugin patterns — but limited strategic value as a live dependency. If the vault needs markdown processing, markdown-it should be evaluated instead. The score reflects a tool that did its job excellently in its era but is now in a maintenance holding pattern with no active development roadmap. The Docusaurus endorsement in the description is partially misleading given that Docusaurus v2+ migrated away from remarkable. Recommend adding to vault as a reference entry only; do not adopt as an active dependency.