Turn a tutorial into a video

A tutorial you have written can be drawn as a film: the board playing the moves, your sentences beside it, and a voice reading them — either a synthesised one or your own.

The film is drawn on the server, not on your machine, so it takes as long as it takes and you can put it aside while it runs. Save the tutorial first; a tutorial that has never been saved has nothing to render.

Say it in your own voice (optional)

In the Tutorial Studio, Record narration opens a screen that walks the tutorial beat by beat while you speak.

Record starts. The screen waits for the microphone first — Waiting for the microphone…
Say what you want about the position on screen, then press Next beat (Space), or the spacebar, to move on. The board plays the move and shows whatever is drawn on it.
Pause and Resume are there for interruptions; Stop ends the recording.
Listen plays back what you made, and Record again replaces it.

Two rules worth knowing. A recording is only usable for a film if it reaches the last beat, and the export says so when it does not. And a recording belongs to the tutorial as it was when you made it — change the tutorial afterwards and the app tells you the recording no longer matches, and offers to record it again.

Export the film

From the studio, or from the row of a tutorial in the Library on Teach, press Export video. The dialog asks a few things:

The sound. One answer of three: your own recording if this device has one — or, for a tutorial made from a recording in Preparation, the voice of that recording — Synthesised voice, or No voice. For a synthesised voice you pick the Language first and then the voice, and Hear this voice reads a sentence aloud so you are not choosing blind.
The words on screen. Comments beside the board — Each sentence is written beside the board as it is read. Turn it off for a film of the board alone.
The size. Higher quality (1080p) — For YouTube or a projector. Slower to render, larger file.

Preview draws a few stills — the opening, the middle and the end — before you spend a whole render on it. That preview is the film's own drawing, so it is what the video will look like.

Export starts it. A bar shows how far it has got and roughly how long is left; Hide puts it away and the render carries on, and Cancel render stops it. When it is done, Download opens the film in your browser, which saves it. The link itself is not shown: for half an hour anyone who has it can download the film, so the button is the only way to it.

The same tutorial in another language

Translate… — on a tutorial's card in the Library, or in the Tutorial Studio under Details… — makes a copy of it in another language. Choose one of the languages it is not already in and press Translate. Its words — the name, the description, the parts' names and every sentence — are sent to a language model, DeepSeek; the moves, arrows and squares never leave the app's server. Every sentence that comes back is checked: a move or a square that is not exactly the one you wrote is sent back once to be done again, and if it is still wrong no copy is made at all. It takes a minute or so, while the app says Translating…

The copy opens in the studio; the tutorial you started from is not changed. In the studio a tutorial with changes not yet saved is not translated — the translation is made of what is saved. The copy has no recording of its own: export it with a Synthesised voice in its language.

Later

A tutorial keeps its last film: its row in the Library on Teach offers Download video without rendering again, or Rendering — show progress while a render is running. Films are not kept for ever — after a while the file is cleared and the app says to export it again.

What can go wrong

The voice list is empty or a voice will not play — That voice could not be played. The server log says why. Choose another voice, or export with No voice.
Somebody else is rendering — one film is drawn at a time, and the bar says how many are in front of yours.
The tutorial is too long — the app refuses before drawing anything rather than halfway through. Split it into two tutorials.
Your recording is not offered — it was made on another device, or it does not reach the last beat.
The recording's voice is not offered on a tutorial made from a recording — a move, a part or a sentence was changed since, so the steps no longer stand where the voice speaks. A corrected word does not do this.