mmdr¶

Fast, native Mermaid diagram rendering for Python — no browser required.

Render Mermaid diagrams directly from Python or the command line. No Node.js, no Puppeteer, no headless Chrome. Just install and render.

pip install mmdr

Quick start¶

CLIPythonJupyter

mmdr -i diagram.mmd -o output.svg
mmdr -i diagram.mmd -o output.png --background '#ffffff'
echo 'flowchart LR; A-->B-->C' | mmdr -i - -o -

import mmdr

d = mmdr.render("flowchart LR; A-->B-->C")

d.svg()                          # str
d.png(background="#ffffff")      # bytes
d.save("output.png", width=1200) # auto-detects format
d.numpy()                        # np.ndarray — no Pillow needed

import mmdr

# just evaluating this in a cell displays the diagram inline
mmdr.render("flowchart LR; A-->B-->C")

Why mmdr?¶

The standard way to render Mermaid in Python is to spin up a headless browser and run the official JavaScript library. It works, but it's slow to start, heavy to install, and awkward to embed in a pipeline.

mmdr takes a different approach: the rendering engine is compiled Rust, shipped as a native Python extension. No JS runtime, no browser, no extra system dependencies — just a ~7 MB wheel.

	mmdr	mermaid-cli
Requires Node.js	❌	✅
Requires a browser	❌	✅
`pip install` only	✅	❌
Startup time	~0ms	~3–5s
Speed	100–1400x faster	baseline

Two backends¶

mmdr ships with two Rust rendering engines in one package:

merman (default) — targets full parity with Mermaid @11.15.0, supports all diagram types
mermaid-rs-renderer — faster and lighter, covers the most common diagram types

mmdr.backends()
# ['merman', 'mermaid-rs-renderer']

See Backends for a full comparison.

Output formats¶

Format	Method
SVG	`.svg()`
PNG	`.png()`
Raw RGBA pixels	`.raw()`
NumPy array	`.numpy()`
File (auto-detect)	`.save("out.svg")`