Project Medusa Open the tool

Layout · typography · prepress · beta

Layout in the browser. Together, all the way to print-ready PDF.

Design multi-page print work, social formats and HTML5 ads together. With real prepress, spot colors, vector PDF and an approval process your clients use without an account.

Open the tool Features Release notes

The tool runs on medusa.id42.xyz. Nothing to install, nothing to set up.

How it works

Variable fonts on the text

Square for width, stem for weight, circle for slant, bookmark for optical size – the other axes as tabs below.

The editor

Type area, master pages, panels – and top right, who else is in the document.

Copy editing and approval

Clients edit text in the document, approve or request changes – without an account.

Prepress report

Images, vectors, fonts, colors, transparency and bleed in detail, with a jump to the object.

Timeline with keyframes

In, out, loop and custom keyframes for offset, size, rotation and opacity.

HTML5 ads

Standard sizes, clickTag, live checks and a ZIP that Google Ads accepts.

Try it yourself →

Three things that matter

Several people in one document at once

Typeset, correct, approve – in real time, with names and cursors, chat and video chat inside the document. One link is enough: clients correct without an account, and nobody sends PDFs back and forth any more.

Print-ready without detours

CMYK, spot colours, overprint, bleed and a vector PDF to PDF/X-4 with a built-in check. Prepress sees resolution, colour space and spacing right on the object.

Professional-grade typography

Kerning from the font, a paragraph composer for calm justification, optical margin alignment, typesetting rules per language – and variable fonts you pull right on the text.

At a glance

20 peoplein the same document at once – measured, with no lost change. With names, cursors, chat and video chat.

A book from partsSeveral documents as one book: continuous page numbers across all parts, one PDF or one per part, a table of contents and an index across the whole book – and a check comparing page size, bleed, print profile and colours of every part.

Change oncelink a logo, an imprint or a parent page from one document – every copy follows

VariantsConditional text: show and hide passages, the type reflows at once

Extensibleyour own commands and panels with a little JavaScript – guide under Help

CMYKspot colours, overprint and real separations

Every axisVariable fonts are pulled right on the text.

25languages for hyphenation, five with spell check

0installs – runs in the browser, offline too

IDMLin and out, with styles, tables and footnotes

HTML5ads for Google Ads, animated, with clickTag

Word ⇄ ExcelTexts and tables stay linked to Word, Excel, CSV and Google Sheets – both ways: changes in the layout travel back to the source.

Open PDFand edit it: shapes, curves, text and images become objects

1800Google Fonts families, one click into the document – as TTF for the print PDF

Translate with AIWhole documents as a language layer or duplicate – free without a key, or with Gemini, Anthropic, OpenAI, DeepL.

Data mergeFrom CSV with images: one PDF, a ZIP per record or a new document to keep working on.

Nine specialities

This used to take several programs, two export detours and a handful of emails. Here it is one document and one link.

HTML5 ads for Google Ads

Pick a standard size, animate on the timeline, set the clickTag, tick off the checklist, export the ZIP. Google accepts it – with no detour through three programs.

WebGL and custom shaders

A surface gets GLSL code, runs immediately and hangs on the same timeline as everything else. It keeps running in the HTML and video export.

Two-way links to Word and Excel

Tables and texts stay connected. Change something in the layout and it travels back to the source file – download or upload a new one included.

Talking right in the document

Whoever is on the same document at the same time sees a speech bubble and a camera in the bottom right. Ask something quickly or look each other in the eye – no second program, no meeting invite, no link to send.

Google Fonts straight into the document

Search around 1800 families, see the real typeface in the preview and install it with one click. The TTF version is fetched – it is the only one that can be embedded in a print PDF.

Approval without an account

Clients get a link, change texts right in the layout, comment and approve. No account, no PDF ping-pong, and the layout stays locked.

PDF/X-4 with a check

The delivery PDF carries the output intent, the embedded print profile and XMP identification. Then the program reads its own file back and checks it: fonts, colour space, page boxes, annotations inside the trim box. Somebody else's PDF can be checked the same way.

Working offline

Once opened, the program starts without a network and installs like an app. Offline documents live in the browser, images and fonts included, and move to the server later at the push of a button. What offline lacks – working together, comments, video chat – is stated where you look for it.

PDF forms

Text fields, checkboxes, dropdowns and signature fields are placed in the layout and end up as real form fields in the PDF – fillable in Acrobat, with the typeface, size and border from the document.

What it looks like

Every image is a real view from the running program, no mockups. Click to enlarge.

**The editor**Type area, master pages, panels – and top right, who else is in the document.

**Prepress report**Images, vectors, fonts, colors, transparency and bleed in detail, with a jump to the object.

**Copy editing and approval**Clients edit text in the document, approve or request changes – without an account.

**Timeline with keyframes**In, out, loop and custom keyframes for offset, size, rotation and opacity.

**HTML5 ads**Standard sizes, clickTag, live checks and a ZIP that Google Ads accepts.

**Three interface styles**The same editor in three styles: «Figma» sober and compact, «Spectrum» with more air, «Studio» dense and grey – light and dark, chosen on first launch, switchable at any time.

**Inspect mode**Everything locked: prepress sees resolution, color space and distances right on the object.

**Format catalogue**49 formats for print, social media, web and video – screen formats with exact pixel sizes.

**Formula setting**LaTeX-style input, preview as you type, vector output in the PDF.

**PDF forms**Text fields, check boxes, lists and buttons right in the layout; the PDF is fillable.

**Creating guides**Columns and rows with gutters, relative to page or type area, colours of your choice.

**Split window**Two views of the same document: detail on the left, the whole spread on the right.

**Linked sources**Excel, CSV, Word and Google Sheets stay connected – two-way if you like.

**Translating with AI**Pick the target language, get a language layer or a duplicate of the document.

**Tidying up a document**The assistant turns identical overrides into styles, merges duplicates and removes what is unused – a preview first, a single undo afterwards.

**Cleaning text**Micro typography by language: quotation marks, dashes, protected spaces – with presets and colour highlighting.

**Print example**Four A4 pages with typesetting, a table, formulas, spot colour, bleed and a form.

**Ad example**300 × 250 px for Google Ads: animation, clickTag and checklist are set up.

**Variable fonts example**Roboto Flex with 13 axes and Source Serif 4 with optical size.

**WebGL example**Your own GLSL shader as a surface, connected to the timeline.

What is inside

Twelve areas with 107 points, and not one of them is merely announced. Open whatever interests you. Project Medusa is a proof of concept: the print PDF is built to PDF/X-4 and checked by the program itself, but there is no independent certification.

Typesetting and typography

Conditional text: give passages conditions and show or hide them; hidden text is neither composed nor exported, and the type reflows at once
Footnotes and endnotes with numbering style, start number, restart per page, section or story, prefix and suffix, their own styles and an adjustable divider rule
Paragraph, character and object styles, nested styles and GREP styles
Justification down to the detail: word and letter spacing, glyph scaling, auto leading, balanced ragged lines, hyphenation with preview
Metric kerning from the font's own pairs – even across format changes –, optical or none; manual with Alt+arrow
Baseline grid, optical margin alignment with a reference size, drop caps, paragraph rules, tabs with leaders
Language detected per text frame: hyphenation in 25 languages, spell checking in five, with Swiss ss
Text frame options: fixed or flexible columns, inset per side, vertical justification, first baseline, custom baseline grid, column rules, auto-size to the text
Bullets and numbering with your own character, number formats, five levels and start number; paragraphs spanning several columns or in a split column
Typesetting rules from Swiss federal, Duden and French typography: narrow non-breaking spaces, quotation marks matching the language as you type, no line starts with a dash
Character attributes as in professional typesetting: horizontal and vertical scale, skew, stroke, character background, underline and strikethrough with weight, offset and colour, true and synthetic small caps, superscript and subscript
Styles with a based-on style and next style, a line style for the first lines, nested styles up to the end of a sentence
Break characters for column, frame, page, odd and even page; keep paragraphs together, equal spacing for the same style, hyphenation exceptions per language

Fonts

Fonts embedded in the PDF as subsets, OTF as CFF – small files, text stays searchable
Browse Google Fonts, preview them and install them into the document with one click – as TTF, so they make it into the PDF
Upload your own TTF and OTF files; variable fonts arrive with all their axes and named instances
OpenType panel with preview: ligatures, figure styles, fractions, numerators and denominators, titling alternates, stylistic sets and character variants – showing what the font can do; plus the glyph panel
Variable fonts right on the text: handles for width, weight, slant and optical size, all further axes as drag tabs – and your own settings saved under a name
Bundled fonts with a free licence, exportable in the package – nothing is loaded from outside servers

Long documents

Linked objects: publish a logo or an imprint once and place it in many documents; when the original changes every copy follows, and deliberate deviations stay
Book: align styles across all parts with a line-by-line preview, plus a table of contents and an index across the whole book
Hide single objects of a parent page per page – no logo on a chapter opener, no folio on a title
Nest parent pages (“based on”) and link them from another document: the copy lives in this document, the reference only reports when the source has changed – output, packaging and offline never need it
Book files: run several documents as a book, continuous page numbers across all parts, one PDF for the whole book or one per part, and a check across all parts
Table of contents with named TOC styles: levels, entry styles, page number before or after, leader characters, alphabetic sorting and PDF bookmarks
Master pages, sections, running headers, text variables
Footnotes, cross references, table of contents and index
Anchored objects, marginalia, captions that follow the image
Tables with continued frames, text wrap, find and replace – with GREP and back references too
Paragraph shading and borders, per column and across frame boundaries

Colour and prepress

Your own preflight profiles: image resolution, thinnest stroke, smallest type, maximum ink, distance to edge and fold – stored in the document and passed on to the whole team as a house-preset file
Preflight test document as an example: 15 pages with spot colours in CMYK, Lab and RGB, overprint, transparency, line art, 16-bit, CMYK and Adobe RGB images – each page states what it tests and what the result must be
Separation and overprint preview straight from the print data: CMYK images keep their channels, RGB images go through their profile, grayscale stays on black only
Spot colours in CMYK, Lab or RGB, on their own plate even in gradients; standard ECI profiles ISO Coated v2, PSO Uncoated and ISO Newspaper included
100 % black overprints for text, fills and strokes; line art stays sharp as 1-bit on the black plate only
Colour picker by colour space: CMYK with four channels and ink coverage, spot colours with tint only, RGB, OKLCH and P3 with a colour field; tint and opacity kept apart; changes apply to the selection only
Inner shadow, feathering per side, global light, isolated and knockout groups – with an Effects window and live preview
CMYK, RGB and spot colors with real separations in the PDF
Color management with ICC profiles, soft proofing, separation preview
Wide-gamut colour models: OKLCH for perceptually even steps and Display P3 for the wide gamut – exported to the web as oklch() and color(display-p3 …) with hex as a fallback
Overprint for objects and black text, plus an overprint preview built from the separate plates
Vector PDF with embedded fonts, bleed, crop marks and slug line
Installable and offline: once opened, the program starts without a network; offline documents live in the browser and can be put on the server later
Built to PDF/X-4: output intent, XMP identification and our own check that reads the finished PDF back – somebody else's PDF can be checked too. It is not an independent certification
Honest labelling: «PDF/X-4» is only written when the file meets it – otherwise the export says why
Drop shadows and feathered edges keep their colour in print: pure black stays black, spot colours stay their own plates; shadows multiply, the object stays vector
Export dialog with presets for the whole team, PDF/X-4 standard, marks with registration marks, bleed per edge, sRGB output and one PDF per page – the PDF is made in the background
Print preview plate by plate on the graphics card: overprint, opacity and spot colours as printed, in Display P3 on suitable screens
Groups with opacity, blend mode, shadow and feathering as a whole; fill, stroke and content each with their own transparency
RGB images are converted to the print profile with their own colour profile, untagged images in the working space; CMYK images stay unchanged
Opacity and blend modes act on the whole object, on screen as in the PDF; soft proof with paper white
Prepress report and inspect mode for sign-off
Prepress disappears once something is animated – an animation is not a printed product

Working together

Real time: several people in the same document, with names and cursors
Chat: short messages on the document, for comment and view links too, with an unread counter
Video chat right in the document: picture and sound go browser to browser, the server only introduces them
Separate links for editing, copy editing, commenting and viewing
Copy editing and approval without an account: change text, comment, approve
Comments on the object, named versions, restore

Motion and web

Timeline with ins and outs, loops and custom keyframes
WebGL surfaces and Lottie animations – they keep running live in the export
Export as video (MP4/WebM) or animated HTML
HTML5 ads for Google Ads with clickTag, checklist and backup image

Formulas, forms, interactive PDF

Formula setting in LaTeX notation: fractions, roots, sums, integrals, Greek letters – vector output in the PDF
Fillable PDF forms: text fields, check boxes, radio buttons, lists, signature, required fields
Buttons with actions: open an address, go to a page, show and hide a field, print, reset
Videos (MP4, WebM) in the layout, embedded in the interactive PDF, still image in the print PDF
Page transitions for full-screen presentation

Data and links

Images from other systems: browse, preview and place from a network folder over WebDAV or HTTP and from Adobe Experience Manager
Links panel: every placed image and PDF with page and status, replace a file in all frames at once; place TIFF, PSD and AI, CMYK stays CMYK
Linked tables and texts from Excel, CSV, Word and Google Sheets – two-way, with write-back and download
Data merge as a PDF, as a ZIP per record or as a new document to keep working on, with a filter
Data merge from CSV, images included
Translation of whole documents with AI, as a language layer or duplicate – free without a key, or with Gemini, Anthropic, OpenAI, DeepL

Production and output

Output automation: workflows, variables in file names, batch output across documents
PDF presets, JPEG and PNG, spreads, booklet imposition
Package as ZIP, IDML export, import of packages and IDML with groups, tables, footnotes and spot colours (Pantone as Lab through the print profile)
Open and edit PDF: shapes, curves, text and images become objects, spot colours become swatches; place a PDF at its original size, link it or insert it editable
Libraries and brand kits for colors, styles, fonts and logos

Tidying up and micro typography

Load styles from another document: swatches, paragraph, character, object, table and cell styles, each marked new, different or identical – only what you tick is taken, in a single undo step
Table and cell styles as named styles, plus object styles with a base style, next style and override marker
Cleaning text: language-aware quotation marks, dashes, measurements, protected spaces – with presets and colour highlighting
Tidy-up assistant: identical overrides become styles, duplicate styles and swatches merge, unused things go – with a preview and a single undo
GREP find and replace with templates for double spaces, dates, phone numbers and more
Saved export presets with an automatic file name suffix (_lowres, _print, _web)

Working and viewing

Folders in the document overview: your own shelves, documents dragged into them. Folders belong to the person, not to the document
Extensions: your own menu commands and panels from a manifest. The code runs in a worker, never in the page – no DOM, no state, no cookies
Layers list as a tree: expand and collapse groups, drag and hide objects inside them; layers with their own colour and «do not print»
Alt-drag duplicates, even mid-drag
Menu bar like on macOS with File, Edit, Layout, Object, Text, Table, View, Window and Help – plus a menu at the document name for everything that concerns the document as a whole
Split window: two views of the same document with their own zoom
Arithmetic in every number field: 10+5, 210/3, 2cm, 12pt
Click a guide and set it to a tenth of a millimetre, put guides in front or behind, lock them, delete one or all
Baseline grid with start, increment and colour – paragraphs follow it with every line or only the first
Show hidden characters: pilcrow, space, non-breaking space and tab
Duplicate by offset or by a fixed gap, align and distribute with a given gap
Duplicate pages and spreads, workspaces, double panel columns, history with a jump to any step
Three interface styles: «Figma» sober and compact, «Spectrum» calm and roomy, «Studio» dense and grey – all light and dark
Interface in German, English, French and Spanish – switch without reloading the document

Accounts, security and feedback

Licences and third-party works under Help: the running version, every notice and the full licence texts, shipped along instead of linked
Self sign-up with an email address, approval by the admins, blocking and password reset on a page of their own
Hardened for teams: throttled sign-in, email confirmation, blocking that takes effect at once, file locks against lost changes, strict security headers
Measured: 20 people in one document at the same time, 1500 requests in 1.3 seconds, no lost change
Ideas, suggestions, bugs and feature wishes: categorised, filterable, with a thumbs up and a counter
Report a bug right in the tool under Help – environment details included
Settings just for you, «Document setup» for everyone: page, grid, colour, interactive – choose interface and appearance on first launch
Delete your own account yourself – with an overview of what disappears and a confirmation you cannot give by accident

Examples you can open

Four ready-made documents wait on the start page – in German and in English. Open them, rebuild them, export them.

Print exampleFour A4 pages

Two-column type area, paragraph and character styles, a table with a header row, two formulas, spot colour HKS 43 K with overprint, an image in the bleed and a fillable form on the last page.

Open the example →

Ad example300 × 250 px, Google Ads

A moving surface in the background, headline and button with ins and outs on the timeline, clickTag set, checklist green – exportable as a ZIP.

Open the example →

WebGL example1080 × 1080 px

A surface with a custom GLSL shader (colour waves) and two animated texts on top. The code can be changed in the dialog and keeps running straight away.

Open the example →

Variable fonts exampleA4 landscape

Roboto Flex with all 13 axes – narrow, wide, slanted, with tall capitals – and Source Serif 4, whose optical size follows the type size. Select text and pull the handles.

Open the example →

Small samples

Arithmetic in a number field210/3 · 10+5 · 2cm · 12pt

Formula in LaTeX notationx_{1,2} = \frac{-b \pm \sqrt{b^2-4ac}}{2a}

Placeholder for the data mergeHallo <<Name>> aus <<Ort>>

Shader linegl_FragColor = vec4(col, 1.0);

Release notes

New in this version2026-09-21

The checkbox is a box again and its label sits next to it instead of at the right edge – in all four languages.
Paste a document link into a chat or a social network and it now shows the document name and its first page instead of a bare address. The link alone is enough; the document itself stays as protected as before, and opening it still asks for a sign-in.
The sign-in page now has «Forgot your password?». You enter your username or email address and receive a link that lets you set a new password within two hours. The link works once; resetting signs out all previous sessions.

All release notes →

Get started

One account, one document, one link – that is all it takes. Whoever has the editing link also sees the links for copy editing, commenting and viewing.

Go to Project Medusa

About me

I am a designer, I have been to a number of Adobe MAX events and I am active in prerelease forums. A prepress professional myself, and I sit on the examination board for Swiss print media technologists.

Why this exists

There is a lot of talk in those forums – but often the people who actually work with these tools every day are missing. So I thought: then I will build a browser-based publishing tool that fixes my own pain points. Just running in the browser is a win in itself.

Jens Riedweg · silo42.design