Meta has released SAM Audio, an open-source AI model that lets users clean up noisy recordings simply by describing what to remove. The tool can isolate voices, music, and background noise using text prompts, visual cues, or time-based instructions, enabling flexible audio editing and sound separation without specialized software. Whether you want to suppress ambient hums, extract a vocal track, or separate music from a mix, SAM Audio handles it through natural-language descriptions and targeted prompts. Because SAM Audio is open source, developers and audio professionals can experiment, customize prompts, and integrate the model into their own workflows. The prompt options include text descriptions of the unwanted noise, visual indications of where edits should apply, and timing cues to focus on specific segments, making SAM Audio suitable for podcast editing, music production, film sound design, and accessibility projects. This article originally appeared on Digital Trends.