Case study
EchoMind
A virtual sound engineer for creators and audio pros. It listens to a mix, suggests changes by genre, mood and intent, and never moves a fader until you say so. Designed as an app, then as a modular hardware device.
Client
Concept product
Role
Product and industrial design
Platform
Website, hardware
Tools
Figma, Blender
Year
2026
Contents
Process
01
Understand
Help, without taking control
02
Design
The website
03
Build
From screen to object
04
Launch
Modules, setups and renders
05
Outcome
What it shows and what is next
EchoMind
Audio / 2026

A sound engineer that suggests, explains and waits. Nothing moves until you accept.
01 Understand
The problem
The brief
EchoMind is a virtual sound engineer: real-time analysis, adaptive mixing and mastering suggestions based on genre, mood and intent. The starting point was an existing, dense VSE mixer interface.
What I looked at
Why creators distrust automatic mixing. Tools that change the sound on their own feel like they are taking the track away from you, so people switch them off.
The decision
Augment, don’t replace. EchoMind proposes, explains why, and waits for a yes. The person stays in control of every move.
Insight
Creators want a second pair of ears, not a second pair of hands.

Website, home

Website, shop
02 Design
The website
Shop
An e-commerce site that sells the system, not a box: the core, six modules and ready-made setups, from studio to podcast.
Product pages
Each setup gets its own page, led by the photography, with the specs, materials and shipping kept short underneath.
Build your own
A configurator where you snap modules onto the grid and the price follows, so the modular idea is something you try, not just read about.
Principle
Show the object first. The shop should feel like picking up the device.

Website, product page

Website, build your own
03 Build
From screen to object
Hardware
Took the same suggest-and-accept loop into a physical device, inspired by compact instruments: a white body, a black screen and a few big keys.
Modular system
Seven modules on a 48 mm grid (core, faders, encoders, speaker, dial, talk and I/O) that snap together with pogo connectors into different setups.
Blueprints and CMF
Every module drawn as a dimensioned blueprint, with an exploded view and a colour, material and finish sheet.
Why it matters
Designing the app and the object together kept one interaction model across both.

Exploded view

Pogo connectors
04 Launch
Showing it
Setups
Gig, Studio and Room setups show how the same modules scale from a pocket rig to a full desk.
Renders
Modelled and rendered in Blender from the blueprints, with the real app UI on the core display.

Home studio

Podcast
05 Outcome
Results

Pack and go
1
Framer template
7
Hardware modules
3
Setups: gig, studio, room
1
Interaction model across both
What I learnt
Trust in AI tools comes from restraint. A suggestion you can see, understand and refuse feels like help.
What is next
A working prototype of the suggest-and-accept mix screen, and a physical model of the core module.
Next project
Maroon →