How to Write Dialogue in Your Script So the Parse Keeps Every Line, With Delivery, Off-Screen and Captions

Parse a short scene written both ways — as lines and as prose — and compare the episode outline: the lines survive verbatim, the prose becomes a summary.
The shape the parse recognises
Speaker: sentence. One sentence per line. Speaker names as they appear in your cast; nicknames are matched when the script introduces them. Brackets between the name and the colon hold the delivery note; brackets inside the line are read as stage direction and not spoken. Screenplay layout (name on its own line, dialogue below) is also read. What the parse does with the rest: Upload a Script and Parse It.

Delivery notes
Put how the line is said in the brackets: (whispering), (through tears), (flat). The note travels with the line into SceneMixer's storyboard and shapes the performance in the clip. If you leave it out, the storyboard adds a plausible one; yours is better. Keep it to a few words; a sentence of direction is acting instruction for the storyboard, not for the line.
Off-screen: phone, radio, voice-over
Write the off-screen marker first in the note: Father (on the phone, hoarse): Come home. The clip then plays the line without showing the speaker’s mouth. Voice-over narration is (voice-over); inner thoughts (inner voice). Without the marker, an absent speaker may be put in the picture. Storyboard side of this: Dialogue Lines, Voice-Over and Captions.
Captions: text the audience reads
Time jumps and place cards go on their own line as (caption: …). On-screen captions are not switched on at the moment, so such a line does not appear as text in the clip, and the video models do not draw readable words on their own. Write time jumps and place cards this way anyway, so they carry over if captions are switched on, and add any readable text in a video editor after download. Text on Screen.
What gets lost
Speech inside narration becomes summary. Lines with two speakers on one line (A: Yes. B: No.) are usually split correctly but are safer as two lines. Very long speeches are kept but will be cut across shots by the storyboard. SceneMixer translates dialogue in a language other than your project’s dialogue language into it at the parse.
| You write | The storyboard gets |
|---|---|
| Mei (quietly): I saw him. | The line, spoken quietly, on Mei's shot |
| She told him quietly that she had seen him. | A summary; nothing spoken |
| Father (on the phone, hoarse): Come home. | The line, heard, speaker not shown |
| (caption: Three years later) | Kept as a caption line; not drawn in the clip for now |
| Crowd: Run! | A group line, spoken in unison |
FAQ
Do I need screenplay formatting?
No. Name-colon lines are enough; SceneMixer also reads screenplay layout.
Can a line have no speaker?
Then it is narration. Give it a speaker or mark it (voice-over) with a narrator name.
How do I mark a line as a whisper?
In the brackets before the colon: (whispering).
Will my exact wording be kept?
Yes, for lines in the recognised shape; SceneMixer's storyboard copies them word for word.
My script is in Spanish but the project speaks English.
Lines are translated at the parse into the project's dialogue language; set that language in Step 1.
One line, one speaker, one sentence
Written that way, the words in the clip are the words you wrote.
Updated November 15, 2026