Preview mode and read-aloud: proofread with your ears

In one line: switch the manuscript to read-only and have it read to you — including a different voice per character in a screenplay.

Why use your ears

By the eighth pass your eye stops seeing your own sentences; it skims. Your ear does not. A missing word, the same adjective three times in a paragraph, a sentence so long you run out of breath — all of that is hard to catch by reading and obvious the moment it is spoken.

There is a practical reason too: read-aloud costs nothing and needs no key of any kind. The voices come from your operating system. Slima only splits the text at its punctuation and hands it over a sentence at a time. No server, no per-use billing, no credits.

Entering preview mode

• A .md on a computer — the view controls at the right end of the tab strip (the cluster next to Sprint, Zen and Split), the book icon.
• A .md on a phone — bottom nav More → Preview mode.
• A .script — the Preview button the screenplay editor already had, on both desktop and phone.

There is no keyboard shortcut. On a computer the button also only appears in the single or left pane, so if you have split the manuscript to the right, switch modes from the left half.

Little changes when you go in: it is the same editor, made read-only. Your cursor and scroll position stay put; the comments rail and linked elements are still there. The one thing that gets replaced is the format toolbar row — the reading toolbar takes it over, so nothing gains height.

Two faces: Reader and Read aloud

The left end of the reading toolbar is a two-way switch.

• Reader (the default) — just read, quietly. No playback controls, and clicking a paragraph does nothing.
• Read aloud — the control row and a hint row appear, and now clicking a paragraph starts speech.

The choice is remembered on this device, so you get the same face next time. Clicking a paragraph only starts speech in Read aloud because of a real complaint: "I clicked the text to edit it and now it jumps around on its own."

What is on the control row

Control	What it does
Play / Pause	One button. Idle plays from the top, speaking pauses, paused resumes
Previous / Next paragraph	Whole paragraphs, not sentences. Next on the last paragraph stops
Stop	Stops, returns to the beginning, clears the highlight
Voice	First option is Automatic, then the voices your device has for this language
Speed	Seven steps: 0.5×, 0.75×, 1×, 1.25×, 1.5×, 1.75×, 2×
Pitch	Five steps: 0.5×, 0.75×, 1×, 1.25×, 1.5×
Cast	Only on a .script — see below

The paragraph being spoken lights up (a tinted background inside a ring) and the page scrolls it to the middle. The highlight moves a paragraph at a time — not sentence by sentence, and not word by word.

Change the voice, speed or pitch mid-flow and it restarts from the current paragraph, because those are fixed at the moment a sentence starts. Switch files or close the tab and speech stops.

Tap a paragraph to start there

In Read aloud, hovering a paragraph draws a thin grey bar to its left: that is the "this paragraph is clickable" cue. Click and reading starts there.

• A click with text already selected is ignored, so copying still works.
• Empty paragraphs cannot be started from.
• Same in a screenplay: whichever line you click is where it starts.

It does not roll on into the next chapter. One file at a time, the one in front of you.

The voices are your computer's voices

Slima ships no voices and uploads no text to synthesise. It asks the browser which voices this machine has, filters them by your interface language, and prefers the ones that run locally — lower latency, and no risk of being cut off part-way through.

So the same manuscript sounds different on different machines. That is normal. Install a better voice in your operating system and it turns up in the list on its own.

Voice, speed, pitch — and the character voices below — are stored on this device and do not follow your account. You set them again on another computer, which is fair enough, since which voices exist depends on the operating system anyway.

A voice per character in a screenplay

A .script gets one extra button on the reading toolbar: Cast. It opens Character voices.

• The list is every character who appears in any .script in the book, most dialogue first, with a line count on each row.
• Each character gets one voice. There is no per-character speed or pitch; those two stay global.
• The default is Narrator (default voice). Scene headings, action, transitions and any unassigned character all read in the narrator voice.
• Parentheticals take the character's voice, because they belong to that character's line.
• Reset all to narrator clears the whole book's assignments in one go.

What gets read is what is on the preview page — the same typesetter that draws the preview and that writes your Word and Fountain exports. Taiwanese format merges a character name with the dialogue that follows into a single line, and that is what you hear: it is not said twice. Lines that carry only format marks (△, ▲) and no real words are skipped.

Watch out

Speech has to be started by you — a browser will not let a page make sound on its own. So there is no "read it to me as soon as I open the file" setting.

Related

• Docs: Zen Mode complete guide
• Docs: Markdown and preview
• Docs: Comments and threads
• Docs: The notes panel
• Docs: Mobile immersive writing
• Docs: Opening a screenplay file