vtt

Documentation

Index

• func FormatAsText(cues []Cue) string
• type Cue
  • func Parse(data []byte) ([]Cue, error)

Functions

func FormatAsText

func FormatAsText(cues []Cue) string

FormatAsText produces a readable "Speaker: text" format suitable for LLM consumption. Adjacent cues from the same speaker are merged into a single line.

Types

type Cue

type Cue struct {
	Speaker string // e.g., "Speaker 1" or empty
	Text    string // cue text content
}

Cue represents a single WebVTT cue with optional speaker attribution.

func Parse

func Parse(data []byte) ([]Cue, error)

Parse extracts cues from WebVTT data. Handles the WEBVTT header, timestamp lines, and <v Speaker N> voice tags. Malformed cues are skipped; returns what can be parsed.