Parse Markdown into a JSON syntax tree.
Free Markdown to JSON (AST) parses your Markdown into a JSON abstract syntax tree — a machine-readable representation of every heading, paragraph, list, code block, table, link, and image. Useful for building static site generators, content pipelines, documentation tools, and any code that needs to understand the structure of Markdown instead of just rendering it.
Privacy: This tool runs entirely in your browser. Your text is never uploaded, logged, or cached. Close the tab and it's gone. Verify in DevTools → Network: zero requests fire.
The parser is open-source (marked). Running it in the browser costs us nothing.
| Input | Output | Typical size | Speed | Login needed |
|---|---|---|---|---|
| Markdown (.md) | JSON AST | Any size | < 50 ms | No |
Here's what this free tool does in detail — every feature is built to solve real problems, runs entirely in your browser, and is free forever.
Uses marked's lexer, the same open-source parser that powers Markdown rendering on GitHub, GitLab, and Dev.to. Full support for GitHub-Flavored Markdown: tables, task lists, strikethrough, fenced code blocks with language hints, and autolinks — all emitted as structured tokens in the output JSON.
To render the same Markdown to HTML instead of JSON, use the Free Markdown to HTML Converter — which uses the same underlying parser to produce human-readable HTML.
Toggle Pretty print to get indented, human-readable JSON with newlines between tokens — perfect for inspection, debugging, or committing to a version-controlled fixtures file. Disable for compact single-line JSON suitable for production pipelines where bytes matter.
The output is valid JSON you can parse with any JSON library in any language — Python, Go, Rust, Ruby, Swift, whatever.
When pretty-print is enabled, pick any indent width from 1 to 8 spaces. The default is 2 (matching most JavaScript and JSON convention), but 4 is also common in some style guides. Tabs aren't offered because tab-width in rendered output depends on the viewer.
For formatting the source Markdown before parsing, use the Free Markdown Formatter — its options include whitespace normalization that can make the AST cleaner.
Every Markdown element becomes a typed token with its complete metadata: heading (with depth 1-6), paragraph, list (with ordered flag), list_item, code (with lang), table (with header, rows, and alignment), link (with href and title), image (with href, alt, title), blockquote, hr, and more.
For extracting just the headings into a table of contents, the Free Markdown TOC Generator does the same traversal and outputs a formatted TOC directly.
The parsing happens inside your browser tab using JavaScript. Your Markdown never leaves your device, never touches any server, never logs. Close the tab and every trace is gone. This makes the tool safe for parsing sensitive documents — NDAs, unpublished drafts, proprietary specs — without worrying about where the text goes.
Same privacy model across the site. The Free HTML to Markdown Converter, Free Markdown Live Preview, and every other client-side tool behave identically.
Click Copy JSON to write the result to your clipboard. Click Download .json to save as a file. Both work on mobile and desktop. Use the JSON in code (parse with JSON.parse()), feed it into content pipelines, or commit it as a fixture for test assertions.
Planning to expose the same parsing via an HTTP API? A public freemarkdowntools API with the same parser is on the Month 4+ roadmap, with a generous free tier.
Paste into the left panel, or click Load example to see a sample.
Toggle pretty-print and indent width to match your needs.
Use it in your code, content pipeline, or test assertions.
Building Hugo, Jekyll, Astro-like generators need the AST to render custom components per node type.
Parsing Markdown docs to generate navigation, anchors, or cross-references.
Transforming Markdown into custom output formats via AST traversal.
Asserting that documentation contains specific structures — headings, links, code blocks — via AST inspection.
Seeing exactly how a given Markdown fragment tokenizes.
Yes. Forever. No watermarks, no limits.
marked's token tree — a compact, practical representation with types like heading, paragraph, list, table, code, link, image, blockquote, and more.
No — mdast (from unified/remark) is a different, stricter format. marked's tokens are simpler and faster. Convert between them if needed.
No. Runs 100% client-side. Nothing is uploaded, logged, or cached.
Yes. Or use the marked library directly in your Node/browser code for the same output. We'll also expose a public API in Month 4+.