# Sherpa Onnx Tts

> Local text-to-speech via sherpa-onnx (offline, no cloud)

- Skill: `foxl-ai/sherpa-onnx-tts` (Agent Skill)
- Install (CLI): `npx skillmds@latest add foxl-ai/sherpa-onnx-tts`
- Raw SKILL.md: https://api.skillmd.com/api/skills/foxl-ai/sherpa-onnx-tts/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: DevOps & Infra
- Author: foxl-ai (https://skillmd.com/u/foxl-ai)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/foxl-ai/sherpa-onnx-tts

---


# sherpa-onnx-tts

Local offline TTS using sherpa-onnx.

## Setup

1. Download runtime for your OS from sherpa-onnx releases
2. Download a voice model (e.g., vits-piper-en_US-lessac-high)
3. Set environment variables:

```bash
export SHERPA_ONNX_RUNTIME_DIR=~/.pilot/tools/sherpa-onnx-tts/runtime
export SHERPA_ONNX_MODEL_DIR=~/.pilot/tools/sherpa-onnx-tts/models/vits-piper-en_US-lessac-high
```

## Usage

```bash
sherpa-onnx-tts -o ./output.wav "Hello from local TTS."
```

## Options

- `-o` / `--output` - Output WAV file
- `--model-file` - Specific model file if multiple exist
- `--tokens-file` - Override tokens file
- `--data-dir` - Override data directory

## Notes

- Completely offline, no API keys needed
- Pick different models from sherpa-onnx tts-models releases
- Supports multiple languages depending on model