ZUO / Technology

The software
behind the presence.

We are bringing conversation, character, memory and physical response into one continuous experience. Each layer has a role in making time with ZUO feel more personal.

Inside the prototype

A little of it, in your hands.

Hear a voice sample, try a touch, or explore a few moments of expression.

01 / See her respond

More ways to connect.

Ready for a touch

Choose a touch region to see the response.

Concept interaction · Generated poses and preset replies.

02 / Hear the voice

A voice to begin with.

An English speech sample generated during prototype testing.

Hey, I’m glad you’re here. How was your day?

Recorded sample · Play, pause or replay.

Five layers. One companion.

01 / Conversation

A conversation with context.

We are designing conversations around what you mean, the thread you are following and the memories you have chosen to keep. The goal is a response that belongs to this moment with you.

Context → Relevant memories → Response

02 / Character & memory

A familiar character. A shared history.

Voice, personality and conversational style shape a recognizable character. Memory adds continuity: preferences, important details and shared experiences, with your permission to save, change or remove them.

Character identity + Memories you approve

03 / Voice & timing

Room to speak. Room to pause.

Natural conversation depends on timing. Our voice work brings listening, spoken responses and turn-taking together, so you can interrupt, take a pause and continue without losing the thread.

Listen → Speak → Pause → Continue

04 / Sensing & movement

Let the body join the conversation.

Touch, gaze and small head movements can give a response physical presence. Our planned interaction layer connects these signals with the conversation, while a separate control layer enforces movement limits and stopping.

Touch or speech → Response intent → Bounded movement

05 / Privacy architecture

Your history, under your control.

We are designing permissions into how conversations, memories and devices connect. You should be able to see what is retained, edit or delete it, and choose what follows your character to another device.

Permission → Storage → Access → Deletion

How ZUO comes together

Building the complete robot.

GOAT’s companion software, the hardware that carries it, and their integration define the complete ZUO product. Software development, hardware work and partner evaluation are planned in parallel. Partner selection and final capabilities remain under evaluation.

  1. 01 / Companion software

    A familiar character, in software.

    GOAT is developing companion AI models and software for character, conversation, approved memory and voice. Together, these elements are designed to give the relationship continuity.

  2. 02 / Hardware and movement

    The body that carries the experience.

    GOAT plans to develop the mainboard, companion computing, interaction electronics and biomimetic skin. Specialist partners would supply the biped body and low-level motion control. Partner arrangements and hardware capabilities still need to be confirmed and validated.

  3. 03 / Integrated ZUO

    Designed to work together.

    GOAT would adapt the mechanical, electrical and software interfaces, then integrate and test how the complete robot responds in use. Each body configuration needs its own integration and validation before its capabilities can be offered as part of ZUO.

Proposed architecture · In development

Interaction & Behavior / Design direction

Bringing each response into a continuous experience.

Connecting voice, gaze, expression and movement into a coordinated response—with room to pause.

  1. 01

    Receive signals

    Bring together speech, touch and user controls to identify the current input.

  2. 02

    Understand the moment

    Use conversation, relevant memories and user settings to decide when to respond.

  3. 03

    Choose a behavior

    Choose an appropriate behavior using intent, permission and available capabilities.

  4. 04

    Coordinate expression

    Coordinate voice, gaze, expression and movement, including interruptions and pauses.

  5. 05

    Confirm and adapt

    Adapt to execution feedback; low-level control limits movement and confirms stopping.

Design direction · The complete behavior system and physical coordination require development and validation.

How memory carries forward / Research notes

Recall that fits the moment

We are exploring how the topic and context of a conversation can guide the recall of relevant details. The goal is to bring forward information that helps this particular exchange.

Context that can evolve

Sources, revision history and shared experiences are planned additions. They would give changing information a traceable context, helping later conversations reflect newer preferences and experiences.

One character across devices

Device pairing and shared identity are planned to connect the physical companion and phone software. The research focus is how an ongoing conversation can carry its context between those settings; cross-device continuity remains in development.

Privacy / Design direction

Data routes and permission checks.

Planned inputs begin with user-enabled microphones and separately authorized sensing. Local saving, online processing and optional cloud sync would remain distinct, with relevant context and visible processing status.

User and device authorization would gate access. Revocation and deletion need checks across affected records and copies; account isolation and the complete privacy center are not yet implemented.

Design direction · Internal software prototype data handling is described separately from this website’s demonstrations.

Internal prototype and data handling

Development status

What works today.
What comes next.

Internal prototype testing covers text conversation, voice transcription, spoken playback and memory confirmation, editing and removal.

Automatic voice interruption, cross-device continuity, the full privacy workflow and physical sensing and control remain in development. The demos above show a recorded voice sample and virtual touch responses; open-ended AI conversation and hardware control are not connected on this website.

Explore the experienceOur privacy approach