Skip to content

Examples

Five runnable scripts demonstrating every voice mode end-to-end.

The repo's examples/ directory contains five runnable scripts:

File Demonstrates
01_basic_tts.py Auto-voice synthesis + sysinfo + audio playback
02_voice_clone.py Voice cloning from a reference WAV (with ASR step)
03_voice_design.py Voice design — three different attribute combos
04_batch_multilingual.py Batch synthesis across 7 languages with mixed modes
05_agent_workflow.py Full agent loaded with every tool, end-to-end

Running

git clone https://github.com/cagataycali/strands-omnivoice
cd strands-omnivoice
python -m venv .venv && source .venv/bin/activate
pip install -e .

# Need an LLM to drive the agent — e.g. Anthropic on Bedrock:
export AWS_BEARER_TOKEN_BEDROCK=...
python examples/01_basic_tts.py

The first run downloads the OmniVoice checkpoint (~1 GB).