valtec-tts
by tronghieuit
DevOps & CloudBrief
Local Vietnamese text-to-speech via VITS2 (offline, no cloud)
Description
What This Skill Does
Local Vietnamese text-to-speech powered by VITS2. Supports 5 built-in speaker voices across Northern and Southern dialects, plus zero-shot voice cloning from a 3-10 second reference audio clip. Runs entirely offline after an initial model download.
Running fully offline means Vietnamese text never leaves the machine, making it viable for private or sensitive content where cloud TTS APIs are not acceptable.
When to Use It
- Generating Vietnamese voiceovers for explainer videos
- Reading Vietnamese documents aloud in assistive workflows
- Cloning a specific speaker's voice for personalized audio content
- Narrating Vietnamese e-learning or audiobook material
- Building offline Vietnamese voice interfaces for local apps
Example Workflow
User asks: Convert this Vietnamese paragraph to speech using the Northern Female voice
- 1
Select speaker NF (Northern Female) from the five built-in voices
- 2
Pass the Vietnamese text to valtec-tts.js with --speaker NF flag
- 3
Specify the output path with the -o flag
- 4
Run local VITS2 inference on CPU or GPU
output.wav file containing synthesized Vietnamese speech in the Northern Female voice
Requirements
- VALTEC_TTS_DIR environment variable set to the cloned repository path
Installation
clawhub install tronghieuit/valtec-ttsRun the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.