The studio

The Studio door of the hub opens a library of documents you write about a language: a note on the verbs of motion, a worksheet for a class, an answer an LLM gave you about a word, typeset. Each document is one Markdown file in the studio’s own dialect, and the studio is everything around that file — the shelf that holds it, the editor that writes it with the typeset page beside it, the page that reads it, and the button that prints it as a PDF whose every word of the target language has been checked.

This section is about the studio as an application: its pages, their buttons and what each one does. What you can write inside a document — target-language runs, lemma headings, boxes, formulas — is the dialect section, and its exercises have a section of their own, Writing exercises.

1. The four pages#

The studio lives at /studio/ and has four pages, each a step of the same work:

The library

(/studio/) lists every document, filters them by language, tag and words, and is where documents come in and go out: + New, Paste LLM answer, Upload .md / .zip, Backup, Load from backup, Download N shown — and Open a file…, which opens a Markdown or zip file where it is, on the computer Parseh runs on, instead of bringing it in (Opening a file from your computer).

The editor

(/studio/doc/<id>/edit, or /studio/new for a document not saved yet) is the Markdown on the left and the typeset page on the right, redrawn as you type.

The document page

(/studio/doc/<id>) is the document as a reader sees it, with its ⚙ panel of typography and the rest, its contents, its glosses as a table and as flashcards, the documents that link to it, its exercises, Build PDF, and Download ▾, which also makes it a page for a website.

The LLM prompt page

(/studio/prompt) is the text you give a language model so that its answer comes back already written in the dialect.

2. A document’s life, in short#

1. Press + New on the library. The editor opens on a starter page in the language the toolbox is set to: a guided tour of everything a document can hold, to keep what you need of and delete the rest.
2. Change the title: line — the title is the document’s name, and two documents of one library cannot share one — and write. The preview on the right follows every keystroke.
3. Save (or Ctrl+S). The document now has a page of its own and a folder for its pictures and recordings.
4. Save & view shows it as a reader sees it. Tag it, open its glosses, answer its exercises, copy them into a deck.
5. Build PDF prints it — at a larger size for a reader with low vision, or in black and white for the photocopier, if you ask.

Or let a model write it: copy the prompt from LLM prompt, ask your question, and bring the .md file it writes back in with Upload .md / .zip.

3. Where the documents live#

Every document is a folder under markdown/library/<language>/, named by an id made from its first title and six random characters (verbs-of-motion-3f9a1c). The folder holds source.md — the Markdown, which is the whole truth about the document — beside meta.json (its tags, dates and build state), images/, audio/ and build/ (its PDF). The id never changes, even when the title does; the folder moves to another language’s shelf only when a save changes the target: line.

markdown/library/  persian/    verbs-of-motion-3f9a1c/      source.md        the document      meta.json        tags, dates, the last PDF build      images/          its pictures      audio/           its recordings      build/           main.tex, main.pdf  english/    ...

You never need to open these folders: the library’s Backup takes the whole of it away as one zip, and Load from backup puts it back.

The same studio, with the same pages and the same buttons, also serves the notes you write beside a book or a video — a second library kept with that book or video: see Notes beside books and videos — and a file you open from anywhere on the computer, as a library of one document in a temporary folder: see Opening a file from your computer.