Drawing from a script#

At a glance#

Drawing from a script - a napkin script says what to draw one instruction a line; napkin-sketch draw turns it into SVG, PNG, PDF, a .skbk or a script that rebuilds it in Adobe Illustrator, with no window, from a shell, from Node, or from any language that can start a process (see Drawing from a script). An AI tool can write the script from a sentence: draw --prompt "a three-box flowchart with arrows".

A napkin script says what to draw in words, one instruction a line, and runs with no window: into the same pages, layers and Bezier anchors the app draws by hand, and out as SVG, PNG, PDF, a .skbk the app opens as editable work, or a script that rebuilds it in Adobe Illustrator.

napkin 1
page 400 300
background #fcfaf5
layer "Card"
color #1f2328 width 3 fill #ffe08a
rect 20 20 360 80 r 12
text "Acme Corp" at 200 70 size 28 align center
link "assets/logo.svg" at 20 120 size 120 120 name "Logo"
rough 0.5
circle 300 200 48

There are three ways in, and all three draw the same bytes from the same script and seed.

From a shell, a batch file or CI, with no window and without loading Electron:

napkin-sketch draw card.napkin --to svg,png,pdf,skbk --out out
napkin-sketch check card.napkin --json    # one line of JSON: the diagnostics, the exit code

From Node, as a dependency:

import { drawSvg } from '@isocialpractice/napkin-sketch';
import { drawFile } from '@isocialpractice/napkin-sketch/node';

const { svg, diagnostics } = drawSvg(text);
const { ok, files } = await drawFile('card.napkin', { out: 'out', formats: ['svg', 'png'] });

From any other language, with the script on standard input and one line of JSON back, as the Python, C, shell and batch callers in docs/api/cli/examples/ do:

subprocess.run([shutil.which("napkin-sketch"), "draw", "-", "--json"], input=script, capture_output=True, encoding="utf-8")

What comes out:

SVG, with the layer tree named three ways, so it opens with its layers in Illustrator and Inkscape, and a linked file kept as its reference.
PNG, from the same rasterizer the composition API uses, with no canvas and no native library; a linked file is drawn from the folder the script is in.
PDF, in vector, every page of a book in one document.
.skbk, the book as the app saves it.
.jsx, a script that, run in Adobe Illustrator, rebuilds the drawing there: its layers as layers and named groups, its marks as paths on their own anchors, its text as text frames, and a linked file placed as a linked file (see Working with other programs).

A mistake in a script is reported, not thrown: card.napkin:7:1: error unknown-verb: ..., exit code 2, and whatever could be drawn is still written. A program that would rather not write the text builds the same script as JSON, the object form, and checks it against the schema. The language is one table, so napkin-sketch verbs lists it, and verbs --json hands all of it to an agent at once.

Or from a sentence, experimentally. napkin-sketch draw --prompt "a three-box flowchart with arrows" has an AI tool write the script first; napkin-sketch checks it, keeps it as a .napkin beside the files, and draws it (see the scripting helper).

Or from a figure. napkin-sketch render hero.skbk --animate walk draws the character on the book's first page through the measured walk cycle: eight frames, hero-walk-1.svg to hero-walk-8.svg, all cut to one box so they play back without shifting, with no AI and no window. The script that drew them is in the JSON report, and Animation Mode's Draw measured frames draws the same in the app (see Animation Mode).

Or from a drawing. In the app, Automate > Generate Script writes the script for a file or for the selected layers, and shows it before you copy, save or open it (see Generating a script).

API.md maps the documentation - ten categories, each with a reference, a quickstart and a cheatsheet - API-QUICKSTART.md is the whole API in five steps, and API-CHEATSHEET.md puts every verb, flag and exit code on one page.