tts-voices

To find the best FlowTTS voice for scenarios like audiobooks or podcasts, the skill reads voices.json and semantically ranks candidates by matching your scene, language, and style, returning the top 3 voice options with metadata like VoiceId and previewUrl.

FlowTTS voice selection supports character narration across five languages: zh, en, ja, yue, and ms. It filters the voice library by these language fields to match your specific narration requirements.

Yes, you can filter FlowTTS voices by gender and tone for customer service. The skill evaluates fields including gender, tone, and scenarios to identify and rank the most suitable voices for your customer service application.

To start selecting a multilingual TTS voice, simply provide your scene description or language preference. The skill intelligently filters the voice library and returns the top 3 matching voice options with their metadata.

When ranking FlowTTS candidates, the skill returns voice metadata including VoiceId, language, gender, and previewUrl. It also provides a brief reason explaining the semantic ranking based on tone, style, and scenarios.

Yes, FlowTTS voice recommendation works without external dependencies. It requires only the local voices/voices.json file to perform semantic ranking and output the appropriate voice metadata for your target scene.