What Scenet does today#

Everything here runs, and every example of it is compiled by the test suite. Each line gives the release it arrived in. Work marked not yet released is on main, and in the documentation of main, but not in what pip install scenet gives you until the next release. What’s new has the detail, release by release.

Each “Written as” names the key, command or function the feature is used through, which the language reference, the command line or the Python API describes in full.

Frontends#

Feature

Written as

Since

A panel, written in YAML

panel, cast, script

0.1

A sequence, each panel saying only what changed

panels, over

0.1

Pages of tiers, columns and insets

pages, page, panels[].insets

0.10

Slanted tiers

tiers[].slant

0.10

A comic script, as writers write it

parse_script()

0.1

Scripts in Dark Horse’s published format, and in French

0.11

A voice from off panel

say.off_panel

0.11

Scripts laid out as a manga script, a screenplay or a letterer’s script

0.13

Characters of your own, as puppet files

load_puppet()

0.1

A JSON Schema for puppet files, for completion in an editor

scenet schema --puppet

0.13

Layout#

Feature

Written as

Since

Shots framed by where they cut the body

camera.shot

0.1

Camera angles

camera.angle

0.1

Figures ordered, overlapped, grounded and looking at each other

staging

0.1

Balloons sized for their words, placed in reading order, with tails

say

0.1

Captions

caption

0.4

Faces and expressions

cast.<id>.expression

0.5

Settings: places, masses, the hour and the weather

setting

0.6

Emanata: plewds, squeans, grawlixes and briffits

cast.<id>.marks

0.8

A supplied picture behind the cast

setting.image

0.12

Figures a supplied picture draws, lettered around

cast.<id>.figure

0.12

Reading right to left, for manga

direction

0.12

Colour, as a twin of the value ladder

palette

0.13

Panels that bleed off the page

panels[].bleed, page.bleed

0.13

Lettering#

Feature

Written as

Since

Lettering drawn as glyph outlines, measured in the face it is drawn in

render()

0.1

Kerning, and emphasis in bold and italic

0.11

A face of your own

lettering.font

0.11

One type size for a page, and small type

lettering.size, say.size

0.11

Sound effects, in the display face

sfx

0.11

The warm caption

caption.tone: warm

0.13

Outputs#

Feature

Written as

Since

An SVG for each panel

scenet build

0.1

A strip of a document’s panels

scenet build --strip

0.1

Panel Core, the resolved layout, as JSON

scenet build --core

0.1

The debug overlay

scenet build --debug

0.1

Lettering as live text

scenet build --live-text

0.1

Output a screen reader reads, with a description of each panel

description, scenet build --no-aria

0.12

A supplied picture embedded, so the SVG stands alone

scenet build --embed-images

0.12

Exports for image models: pose, boxes, depth, mask and semantics

scenet build --conditioning

0.12

A PDF for print, with its trim, its bleed and lettering on the black plate

scenet build --pdf, render_pdf()

0.13

A monochrome edition of a tinted document

scenet build --mono

0.13

Tools#

Feature

Written as

Since

The Python API

compile_source()

0.1

JSON Schemas for panel and scene documents

scenet schema

0.1

The browser playground

0.1

The VS Code extension

0.1

scenet check, its findings as text or SARIF

scenet check

0.3

The checks that compile, --deep

scenet check --deep

0.6

The spec pack, llms.txt and the Agent Skill

0.8

The MCP server

scenet mcp

0.8

A rebuild on every save

scenet build --watch

0.13

Roadmap#

What is planned is in the issues, each scoped to be a release of its own, with what it depends on stated there. What Scenet will not do is said in Why Scenet: it draws no picture with a model, interprets no prose, and leaves style to whoever performs the score.