SAM Audio – Segment any sound using text, visual, or time prompts

Summary: SAM Audio is a unified model that isolates any sound from any source using text, visual clicks, or time spans. It combines speech, music, and sound effect separation into a single promptable system, enabling precise audio segmentation without traditional signal processing.

What it does

SAM Audio separates sounds by interpreting semantic prompts such as text descriptions, video clicks, or time intervals. It replaces multiple specialized tools with one model that understands user intent for audio isolation.

Who it's for

It is designed for developers and creators working on audio-related products requiring flexible and fast sound segmentation.

Why it matters

SAM Audio simplifies audio editing by shifting from technical filtering to intuitive descriptions, streamlining workflows and enabling commercial use.