REV ROBOTICS · AI ROBOTICS PLATFORM

MEET ADALYN.

A voice-controlled AI robot assistant built to bring intelligence, perception and physical interaction into one robotics platform.

Brain
Android · on-device AI
Body
REV Control Hub
Interface
Flutter · mobile + web

01 — Overview

One system, three devices

Adalyn is a robotics platform built around a simple idea: an AI brain and a robot body should be separate machines that speak one protocol. The brain listens, reasons and sees. The body moves, senses and endures. A companion app ties them together — and the whole system can run without the cloud where on-device models are used.

  • Wake-word voice control
  • Multi-provider AI reasoning
  • On-device face recognition
  • Real-time object detection
  • Vector memory
  • Streaming speech output
  • Physical robot control
  • Remote companion app

02 — Architecture

Three devices. One intelligence.

Driver Hub The Brain

An Android application running the AI side of Adalyn on REV Driver Hub hardware or any compatible Android device.

  • Wake-word detection
  • Speech recognition
  • AI provider integrations
  • On-device vision processing
  • Memory integration
  • Streaming speech output
  • Tool execution & robot communication

Control Hub The Body

The REV Control Hub manages everything physical — motors, servos, camera and hardware telemetry — and serves the local robot dashboard over Wi-Fi.

  • Drivetrain, climber & collector
  • Servo pan / tilt gimbal
  • Webcam frame serving
  • Local command execution
  • Thermal & power telemetry
  • Kill-switch safety system
  • Bluetooth audio output

Companion The Interface

A Flutter application for phones, tablets and the browser — the remote experience for pairing, enrollment and everyday interaction.

  • Conversational chat
  • QR pairing & onboarding
  • Face & voice enrollment
  • People / object gallery
  • Settings & provider choice
  • Remote interaction

The Control Hub runs the robot and its local services — internet-based AI processing always happens on the Driver Hub or in the cloud, never on the robot itself.

03 — Capabilities

Built to perceive, reason and act

Voice interaction

A custom wake word starts the loop. Speech is transcribed, answered by the configured AI provider, and spoken back through a streaming text-to-speech pipeline — the first sentence is audible while the rest is still generating.

Sub-second streaming speech

Modular AI providers

Google Gemini by default, with NVIDIA NIM, Groq, OpenRouter and Cloudflare Workers AI available through one provider registry. Adding a provider is a single registry entry.

5 providers · 101 tools

On-device vision

Face detection, multi-object tracking and 512-dimensional face embeddings run locally as ONNX models — recognition does not require a network round-trip.

YuNet · ByteTrack · ArcFace

Object detection

YOLO-class detection identifies objects in the camera stream, with confidence thresholds and fine-grained labels from a second on-device model.

Real-time · on-device

Memory

Conversations, entities and facts persist in a vector store, so Adalyn can recall context across sessions and recognize people it has met before.

Vector · conversational

Physical control

Drive, climb, collect, gimbal and audio output on real REV hardware — behind a kill-switch safety system designed to fail safe.

REV Control Hub

Companion experience

Pairing, enrollment, gallery review and settings from a phone or browser — the same design language as every other Adalyn surface.

Flutter · mobile + web

04 — Engineering

Engineered like hardware meets software

Separation of concerns

The AI brain and the robot body are independent machines joined by a small, explicit protocol — either side can restart without taking the other down.

Android-native AI

Kotlin coroutines, foreground services and ONNX Runtime on the Driver Hub, with performance profiles that adapt to the device it runs on.

Safety by construction

A latching kill flag checked every control loop, physical gamepad semantics, per-subsystem motion timeouts and brake-on-stop motor behavior.

Local-first sensing

Vision and recognition models run on-device, so the robot keeps perceiving when the network does not.

Thermal awareness

The Control Hub supervises its own temperature and drives a cooling fan with hysteresis, smoothing and a fail-safe curve.

One design system

Robot dashboard, phone app and this page are generated from one canonical token set, so the product feels like one product.