VoiceStudio is the open-source, fully-local ElevenLabs alternative — voice cloning, voice design, video dubbing, dictation, transcription & audiobook creation in 646 languages.
Python aiaudiobookcudadubbingelevenlabs-alternativehuggingfacelocal-firstmlxomnivoice-studiospeech-to-texttext-to-speechtranscriptiontranslatettsvoice-aivoice-cloningvoice-generationvoicestudioworkflow
import { describe, it, expect } from 'vitest';
import { parseScript, normalizeSpeaker, parseTaggedScript, hasSpeakerTags } from './parseScript';
const L = '“'; // “
const R = '”'; // ”
describe('normalizeSpeaker', () => {
it('drops a leading "the", trims punctuation, title-cases', () => {
expect(normalizeSpeaker('the fox')).toBe('Fox');
expect(normalizeSpeaker('OWL,')).toBe('Owl');
expect(normalizeSpeaker(' red riding hood ')).toBe('Red Riding Hood');
expect(normalizeSpeaker('')).toBe('Narrator');
});
});
describe('parseScript — screenplay', () => {
it('parses NAME: dialogue lines', () => {
const r = parseScript('FOX: Where are we going?\nOWL: Somewhere safe.');
expect(r).toEqual([
{ speaker: 'Fox', text: 'Where are we going?' },
{ speaker: 'Owl', text: 'Somewhere safe.' },
]);
});
it('does not treat a URL as a speaker', () => {
const r = parseScript('See https://example.com for details.');
expect(r).toEqual([{ speaker: 'Narrator', text: 'See https://example.com for details.' }]);
});
});
describe('parseScript — prose', () => {
it('attributes a quote from a leading dialogue tag', () => {
const r = parseScript('The fox asked, "Where are we going?"');
expect(r.some((s) => s.speaker === 'Fox' && s.text === 'Where are we going?')).toBe(true);
});
it('attributes a quote from a trailing dialogue tag', () => {
const r = parseScript('"Run!" shouted Milo.');
expect(r.some((s) => s.speaker === 'Milo' && s.text === 'Run!')).toBe(true);
});
it('handles curly quotes + "said the owl"', () => {
const r = parseScript(`${L}Somewhere safe,${R} said the owl.`);
expect(r.some((s) => s.speaker === 'Owl' && s.text.startsWith('Somewhere safe'))).toBe(true);
});
it('pure narration → Narrator', () => {
const r = parseScript('It was a dark and stormy night.');
expect(r).toEqual([{ speaker: 'Narrator', text: 'It was a dark and stormy night.' }]);
});
it('unattributed quote falls back to Narrator', () => {
const r = parseScript('A sign read "Keep Out".');
expect(r.some((s) => s.text === 'Keep Out')).toBe(true);
// no name nearby → narrator
expect(r.find((s) => s.text === 'Keep Out').speaker).toBe('Narrator');
});
it('returns [] for empty input', () => {
expect(parseScript('')).toEqual([]);
expect(parseScript(' ')).toEqual([]);
});
});
describe('parseTaggedScript — [Name] tagged scripts (#487)', () => {
it('parses [Name] dialogue lines', () => {
const r = parseTaggedScript('[Alice] Hello there.\n[Bob] Hi, Alice!');
expect(r).toEqual([
{ speaker: 'Alice', text: 'Hello there.' },
{ speaker: 'Bob', text: 'Hi, Alice!' },
]);
});
it('joins multi-line dialogue under one tag until the next tag', () => {
const r = parseTaggedScript('[Alice] First line.\nStill Alice.\n[Bob] Now Bob.');
expect(r).toEqual([
{ speaker: 'Alice', text: 'First line.\nStill Alice.' },
{ speaker: 'Bob', text: 'Now Bob.' },
]);
});
it('attributes prose before the first tag to the Narrator', () => {
const r = parseTaggedScript('Intro narration.\n[Alice] My line.');
expect(r).toEqual([
{ speaker: 'Narrator', text: 'Intro narration.' },
{ speaker: 'Alice', text: 'My line.' },
]);
});
it('normalizes tag names and supports a tag on its own line', () => {
const r = parseTaggedScript('[the FOX]\nWhere are we going?');
expect(r).toEqual([{ speaker: 'Fox', text: 'Where are we going?' }]);
});
it('does NOT treat inline markers as speakers', () => {
// [pause]/[voice:…]/[fast] are synthesis markers, not character tags — they
// stay inside the (Narrator) text, never become a cast member.
const r = parseTaggedScript('[Alice] Wait [pause 500ms] for it [voice:default].');
expect(r).toEqual([{ speaker: 'Alice', text: 'Wait [pause 500ms] for it [voice:default].' }]);
expect(hasSpeakerTags('Just narration [pause] here.\n[voice:x] more')).toBe(false);
});
it('hasSpeakerTags detects only real speaker tags', () => {
expect(hasSpeakerTags('[Alice] hi')).toBe(true);
expect(hasSpeakerTags('FOX: hi')).toBe(false);
expect(hasSpeakerTags('plain prose')).toBe(false);
expect(hasSpeakerTags('[fast] zoom [/fast]')).toBe(false);
});
it('parseScript auto-routes tagged scripts so autoCast just works', () => {
const r = parseScript('[Alice] Hello.\n[Bob] Hi!');
expect(r).toEqual([
{ speaker: 'Alice', text: 'Hello.' },
{ speaker: 'Bob', text: 'Hi!' },
]);
});
});