F5-TTS is an open-source text-to-speech system that achieves state-of-the-art voice cloning quality using a flow-matching approach, enabling high-fidelity voice reproduction from just a few seconds of reference audio. Unlike diffusion-based TTS models that require many inference steps, F5-TTS generates speech in a single forward pass using the Vocos vocoder, making it significantly faster while maintaining exceptional quality. The model excels at preserving speaker characteristics including accent, speaking style, and emotional tone.