md2

Convert markdown files to other formats. Pure Go by default, extensible to new output formats.

Currently supported:

• PDF (.pdf) — syntax-highlighted code blocks
• HTML (.html) — self-contained; local images embedded as data URIs (and remote images too with -flatten); syntax-highlighted code blocks. Diagrams render via inlined mermaid.js (or as static images with -flatten, e.g. for Google Docs import), or as inline SVG for D2 (rendered in-process, no browser)
• Plain text (.txt)
• EPUB (.epub) — EPUB3 ebook (validates with epubcheck). Shares the HTML renderer, so syntax-highlighted code carries over, and the stylesheet has a prefers-color-scheme: dark variant for readers' dark mode. Diagrams (Mermaid, D2, PlantUML) are inlined as SVG in a light and a dark theme, toggled by the reader's color scheme, so they stay legible in both (Mermaid needs a browser at convert time, like the PDF diagram path). A navigation TOC is built from the document's headings; dc:title/dc:creator come from -title/-author. Local images are packaged into the archive

Install

Homebrew (macOS):

brew install rapatao/tap/md2

Nix (flakes):

nix run github:rapatao/md2 -- input.md     # run without installing
nix profile install github:rapatao/md2     # install into your profile

Prebuilt binaries: download the archive for your OS/arch from the latest release. Each release is signed (keyless, via cosign) — see Verifying a release.

Linux packages (.deb / .rpm / .apk): download the package matching your distro and arch from the latest release, then install it:

sudo dpkg -i md2_*_amd64.deb                       # Debian/Ubuntu
sudo rpm -i md2-*.x86_64.rpm                        # Fedora/RHEL/openSUSE
sudo apk add --allow-untrusted md2_*_x86_64.apk    # Alpine

(--allow-untrusted because the .apk is not signed with an apk repo key.)

Go:

go install github.com/rapatao/md2@latest

Or build locally:

go build -o md2 .

Docker (multi-arch, linux/amd64 + linux/arm64, published to GHCR):

md2 reads input from file paths and writes output next to the input (or to -stdout), so mount your working directory and run from it:

docker run --rm -v "$PWD:/work" ghcr.io/rapatao/md2 input.md               # writes input.pdf
docker run --rm -v "$PWD:/work" ghcr.io/rapatao/md2 -f html -stdout input.md > out.html
docker run --rm -v "$PWD:/work" ghcr.io/rapatao/md2 -f html -render mermaid -flatten input.md

The default image bundles Chromium, so every feature works — including browser-fallback PDF, mermaid rendering, and -flatten. A smaller :slim variant omits the browser and covers HTML, txt, D2, PlantUML, and pure-Go PDF only; mermaid, -flatten, and the browser PDF fallback are unavailable there:

docker run --rm -v "$PWD:/work" ghcr.io/rapatao/md2:slim -f html input.md

Tags: latest and <version> (full), slim and <version>-slim (minimal).

Verifying a release

Each release publishes checksums.txt plus a keyless cosign signature (checksums.txt.sig) and certificate (checksums.txt.pem), proving the checksums were produced by the release.yml workflow in this repo (not a tampered fork or mirror).

cosign verify-blob checksums.txt \
  --certificate checksums.txt.pem \
  --signature checksums.txt.sig \
  --certificate-identity-regexp "https://github.com/rapatao/md2/.github/workflows/release.yml@.*" \
  --certificate-oidc-issuer https://token.actions.githubusercontent.com

Then confirm your downloaded archive matches a line in the verified checksums.txt:

sha256sum --ignore-missing -c checksums.txt

Usage

md2 input.md                  # writes input.pdf (default format)
md2 -f html input.md          # writes input.html
md2 -f txt input.md           # writes input.txt (plain text)
md2 -f epub input.md          # writes input.epub (EPUB3 ebook)
md2 -f epub -author "Jane Doe" -title "My Manual" input.md  # set title/author metadata (any format)
md2 -f pdf,html input.md      # writes input.pdf and input.html
md2 -f html -render mermaid -flatten input.md  # self-contained html, diagrams as images (Google Docs)
md2 -f html -render plantuml input.md          # render plantuml diagrams via a PlantUML server
md2 -f html -css extra.css input.md  # append custom CSS after the built-in stylesheet
md2 -o report.pdf input.md    # explicit output (format from extension)
md2 -f html -stdout input.md  # write html to stdout (no file), e.g. to pipe
md2 -f pdf -o book.pdf intro.md chapter1.md chapter2.md  # merge files, in order, into one document
md2 -f pdf -o manual.pdf docs/  # merge every docs/*.md, sorted, into one document
md2 -f html -per-file docs/     # convert each docs/*.md to its own .html
md2 -f pdf -o book.pdf -recursive docs/  # merge docs/ and its sub-folders
curl -s https://example.com/readme.md | md2 -f html -o out.html -  # read markdown from stdin

Pass - as the input to read markdown from stdin (the symmetric complement to -stdout). Because flag stops parsing at the first non-flag argument, put -o/-f before the -. Stdin has no source directory, so an explicit -o (or -stdout) is required, and relative image references resolve against the working directory.

Flags:

• -o output file. Default: the input's name with the format extension. Cannot be combined with multiple formats. Required when merging multiple inputs (several files, or a directory) into one document.
• -f output format(s), comma-separated. Default: inferred from -o extension, else pdf. Duplicates are ignored.
• -render diagram renderer(s) to enable, comma-separated (mermaid, d2, plantuml), or all. Default: none — diagrams render as plain code unless enabled.
• -flatten (HTML only) flatten diagrams to static images instead of inlining mermaid.js, and fetch remote http(s) images and embed them as data URIs, for a fully self-contained file with no JS runtime or external assets needed to view it (e.g. importing into Google Docs). Requires a browser for diagrams, and — new in this flag — network access at convert time for any document that references remote images (so a doc with remote images no longer converts in an airgapped/offline environment under -flatten). A remote image that can't be fetched is left as a live reference with a warning, not a hard failure.
• -user-agent User-Agent header sent when -flatten fetches remote images to embed. Default: a browser-like string, since some hosts reject the default Go client UA. Override for hosts with specific requirements.
• -keep-diagram-source keep the original diagram source in the output in addition to the rendered diagram: the rendered diagram is emitted first, immediately followed by the source as a code block. Default: off — a diagram replaces its source.
• -plantuml-server base URL of the PlantUML server used to render plantuml diagrams to SVG at build time. Default: the public https://www.plantuml.com/plantuml. PlantUML has no pure-Go renderer, so md2 encodes the diagram source and fetches the rendered SVG from this server (inlining it, so the output stays self-contained). This means the diagram source is sent to the server over the network — point it at a self-hosted server for offline or private use.
• -css (HTML output and the browser-rendered PDF fallback only — not the pure-Go PDF path) path to a CSS file whose contents are appended after the built-in stylesheet, so it can override or extend the defaults via normal CSS cascade rules. Local @imports inside it are resolved and inlined recursively (relative to the importing file's directory), so the output stays self-contained; remote @import url(https://...)s are left as-is for the browser to fetch. Since the pure-Go PDF renderer has no CSS support, passing -css with -f pdf forces the headless-browser engine, requiring a browser.
• -per-file with multiple inputs (several files, or a directory) convert each to its own output next to its source instead of merging into one document. Cannot be combined with -o/-stdout (which name a single destination).
• -recursive when the input is a directory, also pick up .md files in sub-directories, ordered folder by folder (a folder's own files first, then its sub-folders). Default: top-level *.md only.
• -author document author metadata: EPUB dc:creator, HTML <meta name="author">, PDF author.
• -title document title metadata (EPUB dc:title, HTML <title>, PDF title); defaults to the document's first heading.
• -stdout write the converted result to standard output instead of a file, for piping into other tools. Single format only. With -o it also writes the file.
• -allow-download authorize downloading Chromium for the browser renderer without prompting (useful in CI).
• -version print the version and exit.

PDF engine

PDF uses a two-stage strategy:

1. Pure Go (goldmark-pdf) — fast, no external runtime. Handles most documents, but has no HTML/CSS layer, so -css has no effect on it.
2. Headless browser fallback — if the pure-Go renderer fails (e.g. complex tables, or glyphs like emoji it cannot lay out), or if -css is passed (custom CSS can only be applied to a rendered HTML document), md2 prints a styled HTML version to PDF with Chrome/Chromium for full fidelity.

The fallback prefers a browser already installed on the system. If none is found it asks before downloading Chromium (~150MB, cached for later runs):

No Chrome/Chromium found. Download Chromium (~150MB) to render the PDF? [y/N]:

On a non-interactive terminal it declines unless -allow-download is passed. Simple documents never launch a browser.

Syntax highlighting

Fenced code blocks with a language tag (e.g. ```go, ```js) are syntax-highlighted automatically — no flag needed — using chroma with the light github theme. Highlighting is applied to HTML (colors inlined as a self-contained stylesheet) and to PDF (both the pure-Go and browser-rendered engines). Blocks with no language, or a language chroma does not recognize, are left as plain code.

Because the HTML colors are emitted as CSS classes, -css can recolor tokens via the cascade — e.g. .chroma .k { color: #b00 } restyles keywords.

Diagrams

Diagram rendering is off by default and enabled per run with -render. Without it, a diagram code block is rendered as plain code.

```mermaid
graph TD; A-->B;
```

```d2
x -> y
```

```plantuml
Alice -> Bob: hello
```

md2 -f html -render mermaid input.md            # enable mermaid (interactive)
md2 -f html -render mermaid -flatten input.md   # diagrams as static images
md2 -f html -render d2      input.md            # enable D2 (inline SVG)
md2 -f html -render plantuml input.md           # enable PlantUML (server-rendered SVG)
md2 -f pdf  -render all     input.md            # enable every supported renderer
md2 -f html -render mermaid -keep-diagram-source input.md  # rendered diagram + its source

Three renderers are supported, with different rendering models:

• mermaid renders client-side. In HTML, the mermaid library is inlined into the output (no network access needed to view it) and the block renders to SVG in the browser — interactive, but needing a JS runtime to display. With -flatten, md2 renders the document in a headless browser and replaces each diagram with a static PNG image, producing a self-contained file that displays anywhere — including a Google Docs import (upload the .html to Drive, then "Open with > Google Docs"), which runs no JavaScript. -flatten also fetches any remote http(s) images and embeds them as data URIs (needing network access at convert time), so the output has no external asset dependencies at all. In PDF, a mermaid block forces the headless-browser engine (the pure-Go renderer cannot run JavaScript), so the diagram is captured as vector graphics. Inlining the library adds ~3 MB to each HTML file that contains a mermaid diagram; -flatten avoids that.
• d2 (D2) renders at conversion time, in-process via the pure-Go D2 library — no JS runtime, no external binary. In HTML the resulting SVG is embedded inline, so the output is self-contained and needs no browser (no -flatten). In PDF, since the pure-Go PDF renderer cannot rasterize SVG, a d2 block routes through the headless-browser engine like mermaid. A d2 block whose source fails to compile is left as a plain code block (with a warning on stderr) rather than failing the conversion.
• plantuml (PlantUML) renders at conversion time by sending the encoded diagram source to a PlantUML server (public by default, or a self-hosted one via -plantuml-server) and inlining the returned SVG, so the output stays self-contained. PlantUML has no pure-Go renderer, so this is the one renderer that reaches over the network — point -plantuml-server at a private instance for offline or confidential diagrams. In HTML the SVG is embedded inline (no browser); in PDF the block routes through the headless-browser engine, as with mermaid and d2, since the pure-Go PDF renderer cannot rasterize SVG.

By default a rendered diagram replaces its source. Pass -keep-diagram-source to keep both: the rendered diagram is emitted first, immediately followed by the original source as a code block.

In plain text output, diagrams are not rendered — the source is kept as code. Files without a diagram are unaffected by any of the above.

The -render flag accepts any combination of the supported renderers, or all to enable every one, and is designed to take additional renderers in the future.

Output naming

When -o is omitted, each output keeps the input's path and base name, swapping the extension for the format — docs/report.md with -f pdf,html produces docs/report.pdf and docs/report.html.

Multiple inputs

Pass more than one markdown file to merge them, in the order given, into a single output document. Merging has no obvious output name, so -o (or -stdout) is required:

md2 -f pdf -o book.pdf intro.md chapter1.md chapter2.md

Flags must come before the input files — Go's flag package stops parsing at the first non-flag argument, so -o/-f cannot follow the file list. Files are concatenated with a blank line between them (so the last line of one file never merges into the first line of the next); heading levels are used as-is, with no automatic page or section break inserted. Each file's relative image references resolve against its own directory.

Pass -per-file instead to convert each input to its own output next to its source, rather than merging:

md2 -f html -per-file intro.md chapter1.md   # writes intro.html and chapter1.html

Directory input

Give a directory as the input to pick up the .md files inside it. The same merge-vs-split rules apply: -o merges every file into one document, -per-file converts each to its own output:

md2 -f pdf -o manual.pdf docs/   # merge all docs/*.md, sorted, into one PDF
md2 -f html -per-file docs/      # one .html per .md, next to each source

Files are picked up top-level only by default; add -recursive to descend into sub-directories. They are ordered folder by folder — a folder's own files first (sorted by name), then each sub-folder in turn, recursively — so naming files 01-intro.md, 02-setup.md gives a predictable merge order. A directory with no .md files is an error. The input must be either a single directory or a list of files, not a mix.

Supported formats

Format	Extension	Engine
pdf	.pdf	goldmark-pdf (pure Go), browser fallback (go-rod)
html	.html	goldmark (GFM), styled standalone document; local images embedded as data URIs; diagrams as mermaid.js (or, with -flatten, static PNGs via go-rod) or in-process D2 inline SVG
txt	.txt	goldmark AST walker, markup stripped, structure kept

Adding a format

Each format lives in its own package under internal/converter/. Create a new one that implements converter.Converter and registers itself in an init:

// internal/converter/docx/docx.go
package docx

import "github.com/rapatao/md2/internal/converter"

type Converter struct{}

func (Converter) Convert(src []byte, w io.Writer) error { /* ... */ }

func init() { converter.Register("docx", Converter{}) }

Then blank-import the package in main.go so its init runs:

_ "github.com/rapatao/md2/internal/converter/docx"

The CLI picks it up automatically — no other changes. It then works standalone (-f docx) and in any comma list (-f pdf,docx).

Layout

main.go                       thin entry point: resolve version, call cli.Run
internal/cli/
  cli.go                      arg parsing, format/render resolution, I/O orchestration
  flags.go                    flag set definition
internal/converter/
  converter.go                Converter interface + format registry
  pdf/pdf.go                  markdown -> PDF: pure-Go first, browser fallback
  html/html.go                markdown -> styled HTML document (+ Render helper)
  html/d2.go                  D2 diagrams -> inline SVG (in-process, pure Go)
  html/plantuml.go            PlantUML diagrams -> inline SVG (via PlantUML server)
  html/assets/                bundled mermaid.min.js inlined into HTML output
  text/text.go                markdown -> plain text (AST walker)
  chrome/chrome.go            HTML -> PDF / diagram capture via headless browser (go-rod)
internal/css/                 -css load, @import inlining, append to stylesheet
internal/merge/               concatenate multiple input files into one document
internal/urlref/              classify image/link references as local paths vs URLs
internal/consent/             interactive prompt gating the Chromium download

Each package has a matching *_test.go alongside it.

Test

go test ./...

License

MIT © rapatao