New slide editor (“Stage”)

• Text (heading, text box) with font, border and alignment
• Image — show whole, fill or stretch; crop; replace
• Video — starts by itself, loop, with sound
• Graphic — can also be broken down to every label (“Ungroup”, “Break apart completely”)
• Shape — rectangle, ellipse, triangle, arrow, line, with fill and corner rounding
• Speaker spot — at most one per slide; frame shape circle, rounded, rectangle or without frame
• Guests — further speakers with their own figure and their own spot (see “Several speakers”)

A slide carries at most 40 elements.

Every step is in the history with an understandable name (“Moved”, “Resized”, “Suggestion accepted”). Named versions save a state permanently. Whoever goes back and does something new does not lose the old path: it is preserved as an “abandoned state” and can be restored.

“Original” shows the slide as it looked when first opened; “Back to template” turns a converted slide back into the original template — this too can be undone.

A generated image or graphic appears in the suggestion window, not on the slide. Several suggestions can be browsed; only “Accept” (or “Replace selected”) inserts it. A guard in the code ensures that no AI output gets onto the slide past this point.

At the core of the editor a slide changes at one single point only: through a command in a plain, storable format, such as “move element 3 by 5 percent to the right”. Mouse, finger and keyboard translate their input into such commands.

This has two consequences: for every state the history also stores the command that led to it, and a second device can replay the same commands and arrives at the same state. New input routes — gesture in front of the camera, phone as a remote control, voice, second screen — therefore need no new core.

The core is a self-contained module without ties to courses, roles or database.

Stage draws a slide with the same building blocks as the course player. A guard checks that editor and player deliver the same rendering — no surprises after publishing.

Older slides with text in fixed pixels also look in the large editor as they do in the player: the editor calculates as if the slide were running in the player. Nothing changes in the course itself.

Triangle and arrow come out of the export as slide deck and as PDF as triangle and arrow, styled words with their styling.

Double-clicking a text opens typing directly on the slide. Above it is a small bar: select words, then bold, italic, underlined, colour, smaller or larger. Without a selection the choice applies to the whole text; “Normal” removes the styling of the selected words.

Older tools continue to see the text as a whole. If they change it, the slide shows the new text plainly — never an outdated one.

“Insert” → “Camera path” opens the stops. On the slide a stop is moved at its edge or its number and dragged larger or smaller at its corners; it keeps the aspect ratio of the slide.

The timeline shows when the camera moves to which point — as in the player at the markers ⟨r⟩ in the narration, without matching markers evenly over the speaking time. “Test run” plays the path (space: pause, arrows: next/previous stop, Esc: close).

Below the slide the bar “Build-up in time with the voice” can be expanded. At the top the narration as a timeline with its markers ⟨r⟩: drag a marker to another position (it jumps to the next word start), double-click sets a new one, × removes it. The spoken text stays the same — finished videos remain valid.

Below it the stages of the elements, as the player plays them: every stage appears at its share of the speaking time. Drag an element into another stage or move it with ‹ ›; “Automatic again” returns to the reading-order sequence.

“Insert” → “Several speakers”: create speakers, choose a figure and give them a spot with “Place spot on the slide”, which can be moved and scaled on the slide. A speaker with their own figure is a guest; they stand at their spot in the player and in the presentation.

Under “Who speaks which sentence” every sentence of the narration is assigned to a speaker. “Videos per section” shows the sequence of sections with the status of their videos; there the selected sections are ordered on request (see “Several speakers on one slide” under “Workshop”).

Every face speaks only with its own voice. As long as not every section of a slide has a finished video, the course's speaker speaks the whole slide, and a guest appears as a still image at their spot. Finished videos are never re-created automatically.

History and saved versions of a slide are also saved on the server. So you find on the tablet what you saved on the laptop; the more recent state applies, nothing of the older one is lost (“State from another device”). If the corresponding database extension is missing on an installation, the history stays in the device's browser.

The bar “Slide deck” lists all slides of the course, organised into chapters; every chapter can be collapsed. “Chapter from here” on a slide starts a chapter, a double-click on the name renames it, the arrows move it together with its slides (within a module), × dissolves it — the slides remain. The arrows on the first and last slide of a chapter pass it to the neighbouring chapter.

“Suggest chapters” reads the modules or the slide titles (numbered titles, the same prefix such as “GDPR: …”) and shows a suggestion. Only “Accept” creates the chapters.

“Insert table of contents” creates a slide with the chapters (or slides, or both) and their slide numbers. It is updated with every change to titles, chapters and order and marked as “Contents” in the list; heading, font and colour stay as the author sets them. “Detach” turns it into an ordinary slide. With up to four entries it gets jump buttons to the chapters in the course; in the export (slide deck, PDF) it appears with text and numbers.

The presentation does not jump when an entry is clicked; there you page through as usual. The numbers count the whole course — whoever exports only individual modules sees the course's numbers.

Search in slide text, titles, narration and intent/notes — each selectable —, with match case and “whole words only”, in all slides, one chapter or the ticked slides. The hit list shows slide and context; “Replace” replaces one hit, “Replace all” all of them in one step, which “Undo” in the bar reverses completely.

Styled words keep their styling, the markers ⟨r⟩ in the narration stay where they are, and slides from templates remain templates. If a narration changes, finished videos are not regenerated; a recording with old text appears as outdated under “Publish”.

Below the design bar is “Apply to other slides”. What: the last design change (e.g. colour of the title — it then applies to the title of every slide, likewise for text, header and footer, speaker, image, shape), the entire design of the selected element, the background, the speaker spot, the logo or the transition. Where: only this slide, selected slides, whole chapter or all slides. Beforehand it says how many slides are affected.

Slides that deviate by hand remain unchanged unless “Also overwrite deviating ones” is ticked. Slides from templates without a matching element also stay as they are. Every application is one step and can be undone in the bar “Slide deck”.

• Types: original, saved version, abandoned state, suggestion.
• Steps with names, including: Moved, Resized, Rotated, Design changed, Aligned, Distributed, Layer changed, Inserted, Deleted, Duplicated, Grouped, Text changed, Graphic broken apart, Image replaced, Suggestion accepted, Beam photo, Figure chosen, Speaker changed, Camera path changed, Build-up changed, Changed from outside.
• “Restore” brings back a state, “Delete version” removes a saved version.

Every author of the course in the workshop. If the feature “New slide editor” is off for the tenant, “Edit large” opens the previous editor of the free surface.