Markdown to HTML Converter
Turn a Markdown draft into HTML for a CMS or a standalone page. Inspect the code, review conversion notices, and check a safe preview before you export.
Conversion runs in this browser. Raw HTML is escaped, images become labelled text, and preview links are inactive. This tool supports the documented subset below, not all CommonMark or GitHub Markdown.
Up to 100,000 UTF-16 code units, 3,000 lines, and 10,000 code units per line. Emoji can use two code units.
.md, .markdown, or .txt, up to 400,000 bytes. Import replaces the source.
Output is plain text here. It is never inserted into the page as HTML.
Keyboard shortcut: Ctrl+Enter or Cmd+Enter while editing.
Ready. Paste Markdown or load the example, then convert.
Review these conversion notes
Safe preview
Shows the document body. Links are inactive, images are text, and no embedded resource loads. Your website's CSS may look different.
Convert, inspect, then publish
- Paste your draft or import a UTF-8 Markdown file. Use blank lines between paragraphs and explicit indentation for nested content.
- Choose a fragment for a CMS content field, or a complete document for a standalone .html file. Shift headings when the surrounding page already has a title.
- Select Convert to HTML. Review the notes and preview, then copy or download the HTML. Editing any input clears the previous result so you cannot accidentally export stale content.
Supported Markdown
Headings with # or setext underlines, paragraphs, thematic rules, fenced and indented code, explicitly marked blockquotes, ordered and unordered lists with indented continuation lines, nested lists, emphasis, strong emphasis, inline code, inline links with optional quoted titles, and HTTP(S) angle autolinks. Pipe tables, two-tilde strikethrough and task markers are also supported. Task markers export as readable [Done] or [To do] text.
Two trailing ASCII spaces or a backslash make a hard line break. Ordinary newlines stay soft, and non-ASCII spaces remain part of the content. Tables need a header and matching delimiter row; escape a pipe as \|. A body row without pipes fills the first cell, with remaining cells empty. Use a blank line to end a table before a paragraph. Extra body cells are omitted with a note. Code is escaped and preserved without syntax highlighting. Nested containers are limited to 16 levels.
Deliberate differences and limits
This is a practical subset converter, not a full CommonMark or GFM implementation. Reference links, footnotes, math, diagrams, automatic bare-URL linking, inline HTML, and lazy unindented list or quote continuation are not supported. Keep every quote line prefixed with > and indent list continuation lines to the item's content column. Align sibling list markers exactly. Blank lines make enclosing lists use paragraph wrappers, so nested-list spacing may differ from other renderers. Unsupported syntax usually stays visible as text; notices are helpful checks, not a complete Markdown validator.
Raw HTML remains escaped text. Image syntax becomes [Image: alt text] without an image URL. Named entities supported are amp, lt, gt, quot, apos and nbsp; numeric entities are decoded. Other named entities stay literal and trigger a note. Tabs expand to four-column stops outside top-level fenced code, which can affect unusual inline tab spacing. Unclosed code fences continue to the end and trigger a note. Generated formatting is limited to 64 nested HTML elements.
Links allow http, https, mailto, fragments, query strings and relative paths. Protocol-relative URLs, credentials, whitespace, control characters, malformed escapes, unresolved entities and other schemes are blocked. Percent-encoded control characters and colons are conservatively rejected. HTTP(S) URLs use the browser's standard URL normalization. Link titles must use single or double quotes. Link checks do not verify destinations or their content. Raw markup and images are never activated, including in exported HTML. A complete document contains no scripts, stylesheets, external fonts or automatically loaded resources; its language is left unspecified.
Privacy and standards
The converter does not upload, store or send your Markdown. Import, preview and downloads run locally; reloading the page clears the tool's in-memory draft. Normal shared site analytics may still record the page visit. No remote conversion service, URL fetch, account or API key is used.
Parsing choices are informed by CommonMark 0.31.2 and the GitHub Flavored Markdown specification. Review the generated fragment in your destination before publishing, especially when its Markdown dialect differs.