Gå til hovedinnhold

WebVTT format (.vtt)

WebVTT (Web Video Text Tracks) is the W3C subtitle and caption format used by HTML5 <track> elements. A .vtt file opens with a WEBVTT header and then lists cues: an optional cue identifier, a timecode line (HH:MM:SS.mmm --> HH:MM:SS.mmm plus optional positioning settings), and one or more lines of cue text.

The reader emits the WEBVTT header as a non-translatable data part, each cue identifier as its own data part, and one translatable block per cue. The timecode (including any cue settings) is preserved on the block's timecode property and written back verbatim, so the timing survives the round-trip. Cues are kept separate — captions are not merged by sentence — so each cue is translated independently.

IDvtt
SourceBuilt-in
Extensions.vtt
MIME Typestext/vtt
CapabilitiesRead + Write

Parameters

ParameterTypeDefaultDescription
cjkCharsPerLineinteger0Maximum number of characters per line for CJK languages. Set to 0 to disable.
maxCharsPerLineinteger0Maximum number of characters per line in output. Set to 0 to disable.
maxLinesPerCaptioninteger0Maximum number of lines per caption in output. Set to 0 to disable.
splitWordsbooleanfalseSplit words so that they don't go over the character limit per line.

Configure these parameters interactively and copy the YAML on the Format Reference.

Examples

Default extraction

Read each cue as a translatable block with no output line limits.

{}

Processing notes

  • The WEBVTT header is emitted as a data part; cue identifiers become their own data parts.

  • The cue timecode and its settings are preserved on the block timecode property and written back unchanged.

Limitations

  • The output line-wrapping parameters (maxCharsPerLine, maxLinesPerCaption, cjkCharsPerLine, splitWords) are accepted for parity with Okapi but are not yet enforced by the native writer; cue text is written as supplied.

  • Cues are never merged by sentence; each cue is a separate block.

← Back to the Format Reference