remove-markdown

remove-markdown is a small, dependency-free Node.js module that strips Markdown syntax out of a string, leaving behind plain, readable text. It targets the full range of common Markdown constructs — headers, emphasis, links, images, blockquotes, list markers, fenced code blocks, GitHub-Flavored Markdown strikethrough, and embedded HTML tags — collapsing them into their underlying text content rather than deleting it outright.

The library is commonly reached for when an application needs to show a Markdown-authored post, comment, or note somewhere that can’t (or shouldn’t) render Markdown: list previews, search result snippets, push notifications, RSS feed summaries, or plain-text email digests. It ships a small options object for tuning behavior — replacing links with their URL, preserving specific HTML tags, or customizing list-leader handling — and a bundled TypeScript declaration file for typed usage out of the box.

What You Get

• A single synchronous removeMd(markdown, options) function with zero runtime dependencies to install or configure
• Coverage for the common Markdown surface: ATX and setext headers, emphasis/strong, blockquotes, list leaders, horizontal rules, inline and fenced code, images, links, footnotes, and reference-style links
• GitHub-Flavored Markdown extras out of the box — strikethrough (~~text~~) and fenced code block stripping — toggleable via the gfm option
• Fine-grained options: useImgAltText to keep image alt text, replaceLinksWithURL and separateLinksAndTexts to control link output, htmlTagsToSkip to allowlist specific HTML tags instead of stripping them, and throwError to opt into exceptions instead of silent fallback
• A bundled index.d.ts TypeScript declaration, so consumers get typed options and return values without an extra @types package

Common Use Cases

• Generating a plain-text excerpt or teaser from a Markdown blog post for a listing page or card
• Producing clean summaries for search indexes (Algolia, Elasticsearch, Meilisearch) from Markdown-authored content so results don’t show raw syntax
• Building plain-text push notifications, SMS, or email digests from Markdown source (e.g. a comment or changelog entry)
• Sanitizing Markdown-formatted user input before feeding it to text-only surfaces like meta descriptions, social share previews, or voice assistants

Under The Hood

Architecture remove-markdown is a single exported function, module.exports = function(md, options), defined entirely in index.js. It runs the input string through a deliberately ordered chain of String.replace() calls: horizontal rules are stripped first (explicitly ordered ahead of list-leader stripping to avoid a regex conflict noted in a code comment), then list leaders, GFM constructs (setext headers, strikethrough, fenced code), abbreviations, HTML tags (built dynamically as a negative-lookahead regex when htmlTagsToSkip is set), links/images/footnotes/blockquotes, atx-style headers, emphasis markers, and inline code. The whole transform chain is wrapped in a single try/catch — on error it either rethrows (when options.throwError is set) or logs to console.error and returns the original, unmodified input, giving callers a safe-by-default fallback rather than a thrown exception. There is no internal module boundary, class hierarchy, or state; execution is a single synchronous pass with no async paths.

Tech Stack The library is plain, dependency-free JavaScript (CommonJS module.exports) with no runtime dependencies at all — package.json lists only devDependencies (chai, mocha, should) used for testing. There is no build step or bundler: the file published to npm (index.js) is the exact source that executes, and TypeScript support is provided by a separately hand-maintained index.d.ts rather than being generated from typed source, so the two can in principle drift out of sync.

Code Quality test/remove-markdown.js contains 264 lines of Mocha/Chai specs covering list stripping, HTML stripping, headers, emphasis, links, images, blockquotes, and the htmlTagsToSkip option, including edge cases like unmatched markdown characters — solid behavioral coverage for a single-function library. There is no linter configuration in the repo and no TypeScript source (only the hand-written declaration file), so type safety is external to the implementation. Naming is terse and regex-driven, which suits the file’s narrow, single-purpose scope; error handling is intentional rather than incidental, via the try/catch plus configurable throwError.

API Design The public surface is a single function, removeMd(markdown, options), which is about as low-boilerplate as an integration gets — const removeMd = require('remove-markdown') and one call. Sensible defaults (gfm: true, stripListLeaders: true, useImgAltText: true) mean zero-config usage already covers the common case, while nine well-documented options in the README cover the less common ones. Naming across options is slightly inconsistent (stripListLeaders vs. useImgAltText vs. abbr), and interactions between options like replaceLinksWithURL and separateLinksAndTexts aren’t explicitly documented, but the overall surface remains small and easy to pick up.