Sound and listening systems for the agentic era.

Phonostack builds infrastructure for sound in a world where computation listens, speaks, generates, edits, organizes, and performs.

Sound is no longer made only by hands. It is now also generated, interpreted, organized, and transformed by humans, agents, models, and machines. We build the tools that hold this new reality together.

01
Sound is infrastructure.
Not a feature, not an asset. The substrate on which agents, models, and humans now act.
02
Listening precedes generation.
A system that cannot listen cannot decide. We treat audition as a first-class capability.
03
Material before performance.
Sound is matter. Granular, spectral, plastic. Tools should expose its interior, not just play it back.
04
Memory over output.
Every generation, prompt, and gesture is held. The library is alive, navigable, and shared.

Sound

Material substance

Listening

Computational perception

Agentic

Human · model · machine

Generative

Emergence from systems

Memory

Sonic cartography

Stack

Layered composition

A running record of releases, experiments, and shifts. Newest first.

• 2026.05Launchphonostack.com goes liveThe ecosystem has a home. Six tools, one listening system.
• 2026.05StudioPhonostack Studio — beta opensGenerative sound design, prompt-to-stack, ElevenLabs-powered.
• 2026.04ORAMv0.1 — open-sourcedAgentic looper for sonic discovery and live performance.
• 2026.04OriginPhonostack foundedA Momoto Studio spinoff, dedicated entirely to sound and listening.