sAtIe is a human-in-the-loop system for authoring generative spatial audio. It is a creativity-support tool that allows sound designers to sequence, position, and transform audio using an expressive domain-specific language that works alongside natural-language prompting. A generative pipeline can produce an initial spatial audio scene from a prompt, after which sAtIe provides a concise and interpretable syntax for shaping timing, variation, procedural movement, and spatialization.

By giving creators an explicit structural layer they can inspect and modify, sAtIe preserves creative intent across iterations and keeps humans in control of the generative process. This enables faster experimentation, more precise adjustments to complex scenes, and tighter alignment between artistic intention and system behavior.