Technology / Platform architecture

Three layers. One reusable robotics architecture.

Partners configure the mission; the patented platform supplies the intelligence and physical foundation.

APPLICATION LAYER

Branded workflows, market logic, integrations and customer-specific user experience.

ROCK OS + EDGE AI

Autonomous navigation, speech, perception, task orchestration, device and fleet interfaces.

SMART CONNECTOR + MODULAR HARDWARE

Validated base, head, torso, sensor and payload configurations with system-level controls.

Technology / ROCK architecture

One protected core.
A configurable client application.

ROCK supplies the shared robotics functions; sector and customer software determines what the robot actually does.

ROCK CORE | CASSINI IP

  • • Robot control, perception and navigation
  • • Device and sensor access through APIs
  • • Voice, identity and task interfaces
  • • Security, updates and fleet management

APPLICATION LAYER | PARTNER IP

  • • Hospital rounds or medication workflow
  • • Retail, hotel, vending or facility tasks
  • • Customer branding, screens and processes
  • • Reusable vertical software licences

White-label architecture lets a partner own customer applications and relationships while retaining

Technology / Offline AI architecture

Fully autonomous in normal operation -
connectivity is optional.

Core AI functions are processed on the robot. Connectivity can be used for setup, updates, fleet services and customer-system integration.

  1. 01

    Initialize

    Install approved software, models, maps and customer configuration.

  2. 02

    Operate locally

    Navigation, recognition, speech and control are executed on the device.

  3. 03

    Sync when available

    Updates, reporting or fleet coordination can be synchronized when enabled.

No mandatory cloud inference charges for the on-device functions; standard maintenance, electricity, support, update infrastructure and customer integrations still create operating costs.

Technology / Offline AI economics

A different scaling curve from cloud-dependent AI.

Commercial benefits derive from running the routine perception and control workload locally.

COST /
OPERATING FACTOR
CONTINUOUS CLOUD - DEPENDENT MODELCASSINI EDGE-AI
MODEL
Inference costsCan increase with interaction volumeNo mandatory per-interaction cloud AI inference
ConnectivityNetwork availability can determine core service uptimeCore functions work without a continuous connection
Data handlingMay need external processing depending on architectureLocal processing and storage supported
Commercial pricingUsage charges may complicate fixed pricingPredictable per-device processing costs

Offline AI creates an opportunity for stable fleet economics

Technology / Standard functionality

Robotic Intelligence and autonomy
come standard.

Capabilities offered by Physical AI-powered devices; role-specific features depend on the configuration and customer environment.

Core offline operation

Local control and AI inference without continuous internet or cloud access.

Mapping & navigation

Real-time mapping, obstacle avoidance and support for pre-ingested maps.

Identity & recognition

Face, object, barcode and QR-code recognition for controlled workflows.

Natural interaction

Voice recognition, dictation and multilingual communication.

Security & traceability

Encrypted local storage, recordings and configurable audit trails.

Modular integration

Sensor, payload and application interfaces designed for customer-specific workflows.

Scout and Cassi robots beside an RMS-India drone carrier at an exhibition stand