What it does

This skill enables high-fidelity voice cloning and text-to-speech (TTS) generation directly through your AI agent. By leveraging the NVIDIA Magpie TTS NIM, it can replicate any voice from a brief 10-30 second audio sample or generate professional narration using high-quality preset voices.

Why use this skill

Integrating professional-grade voice synthesis into a developer workflow usually requires complex SDKs or expensive subscriptions. This skill streamlines the process by using NVIDIA's zero-shot cloning technology, allowing your agent to produce localized audio assets, narration, or personalized voice feedback without leaving the terminal. It is significantly faster than manual audio processing and utilizes a powerful cloud infrastructure for low-latency synthesis.

Supported tools & features

• NVIDIA Magpie TTS Zeroshot: Clone voices from WAV/MP3 files with minimal data.
• NVIDIA Magpie Multilingual: Support for diverse accents and languages including Spanish, French, and German.
• Local File Management: Automatically manages audio output and storage in a dedicated local directory.
• Bypass Setup: Works with a simple API key, removing the need for local GPU-heavy TTS models.

Output format

The skill produces high-fidelity WAV audio files stored locally, providing clear, natural-sounding speech that is ready for use in applications, videos, or testing.