vibevoice
por javier887
Web & Frontend DevelopmentResumen
Local Spanish TTS using Microsoft VibeVoice.
Descripción
Qué hace esta skill
Local text-to-speech tool using Microsoft's VibeVoice model. Converts text to natural Spanish (or English) audio, outputting files in formats compatible with WhatsApp voice messages. Runs entirely on local hardware with no cloud API required.
Runs locally with no API key or internet connection, so there are no per-request costs and audio stays on your machine.
Cuándo usarla
- Send WhatsApp voice messages without recording your voice
- Convert long Spanish text to audio for review
- Generate voice audio files from text scripts
- Test different Spanish or English voice styles locally
- Create audio output from text files in batch
Flujo de trabajo de ejemplo
User asks: send a WhatsApp voice message to +34600000000 saying 'Nos vemos mañana a las 10'
- 1
Run vv.sh with the target text and output path set to /tmp/mensaje.ogg
- 2
Wait for model to generate the audio file
- 3
Verify the .ogg file was created at the expected path
- 4
Send the file via the message tool with channel=whatsapp and asVoice=true
WhatsApp voice message delivered to the specified number
Requisitos
- NVIDIA GPU with ~2GB VRAM
- VibeVoice repository cloned and installed at ~/VibeVoice
Instalación
clawhub install javier887/vibevoiceEjecuta el comando de la CLI o pega el prompt en Claude Code, Cursor, Codex o el agente de programación que uses.