Project Files

mock

text_preprocessor.py

schemas

formatting.schema.json

prepare.schema.json

src

formatter.ts

index.ts

language-packs.ts

narrative.ts

narrator.ts

preparer.ts

processor.ts

types.ts

INSTRUCTIONS.md

manifest.json

package-lock.json

package.json

README.md

tsconfig.json

schemas / prepare.schema.json

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "$id": "https://primary-data.dev/schemas/vibevoice-tts/prepare.json",
  "title": "VibeVoice TTS Prepare Options",
  "description": "Describes the options for the vibevoice-tts prepare tool — the full podcast script preparation pipeline. Handles abbreviation expansion, pause markers, direction tags, paragraph splitting, script chunking, and TTS parameter recommendations.",
  "type": "object",
  "properties": {
    "text": {
      "type": "string",
      "description": "The raw podcast script text to prepare. Supports plain text, Speaker N: format, pause markers like [pause 1s], and direction tags like [emphasis].",
      "minLength": 1
    },
    "chunkSize": {
      "type": "integer",
      "description": "Target words per chunk for incremental TTS generation. 500 words is optimal for VibeVoice quality.",
      "minimum": 100,
      "maximum": 2000,
      "default": 500
    },
    "paragraphPauses": {
      "type": "boolean",
      "description": "Insert empty-line pauses between paragraphs. VibeVoice interprets empty lines as natural pauses.",
      "default": true
    },
    "sentencePauses": {
      "type": "boolean",
      "description": "Split long paragraphs into shorter ones (max sentences per paragraph). Improves TTS quality for dense text.",
      "default": false
    },
    "maxSentencesPerParagraph": {
      "type": "integer",
      "description": "Maximum sentences per paragraph before auto-splitting. Only used when sentencePauses=true.",
      "minimum": 1,
      "maximum": 10,
      "default": 3
    },
    "expandAbbreviations": {
      "type": "boolean",
      "description": "Expand common abbreviations (e.g. becomes for example, AI becomes A I). First occurrence only. Improves TTS pronunciation.",
      "default": true
    },
    "preserveDirections": {
      "type": "boolean",
      "description": "Extract direction tags ([emphasis], [slower], etc.) as metadata hints instead of stripping them. Guides TTS param adjustments.",
      "default": true
    },
    "output": {
      "type": "string",
      "description": "Output format: 'plain' (prepared text), 'script' (Speaker N: prefix format), or 'json' (structured chunk array).",
      "enum": ["plain", "script", "json"],
      "default": "script"
    }
  },
  "required": ["text"],
  "additionalProperties": false
}