Captions

Generated from project.v1.schema.json at the documented engine version. The schema checks shape only; the engine’s validation also checks ranges, limits, references and keyframe order.

CaptionCue

Text shown over one timeline interval.

Field	Type	Description
speaker	string | null	Who says it, as the source names them (a WebVTT voice), at most 64 bytes. A track style of that name applies when the cue names no style.
style	string | null	The name of one of the track's styles, applied over its style.
textrequired	TextContent	The text, a string or runs; \n starts a line. Non-empty, at most MAX_TEXT_CONTENT_BYTES bytes. Run times are from the cue's start and lie within it.
timeline_rangerequired	TimeRange	When the cue shows. Non-empty and not before zero.

Captions

What a caption track shows: its cues, in its style.

Field	Type	Description
cuesrequired	CaptionCue[]	Cues in timeline order, none overlapping; a track shows one cue at a time.
style	TextStyle	The style every cue starts from. Default: {"align":"center","anchor":{"x":0.5,"y":1},"background":n…
styles	{ [name]: TextStylePatch }	Named changes to style for cues that name one or whose speaker has one. At most MAX_CAPTION_STYLES, names 1 to 64 bytes.