Develop, connect and deploy IoT devices and AI applications on one communication layer, from a single microcontroller to an industrial sensor network.
A high-availability communication stack for IoT devices, Edge AI accelerators and autonomous robotics. Connect your own AI and LLM models directly to the network layer, and route telemetry through a SQL rule engine and managed MCP servers for real-time orchestration between intelligence and hardware. Manage everything from low-power microcontrollers to robotic fleets in one project environment.
Technical components for deploying and managing distributed device ecosystems.
High-scale messaging with support for MQTT v3.1.1/v5.0, MQTT-SN, and WebSockets.
TLS 1.3 encryption, mutual authentication (mTLS), and granular Access Control Lists (ACL) for topic-level permission management.
Real-time data processing using SQL-like syntax. Extract, filter, and route telemetry to Kafka, webhooks, or cloud service providers.
Communication layer for orchestrating data between microcontrollers, edge accelerators, and autonomous systems.
Modern IoT communication protocols via MQTT, MQTT-SN, and WebSockets.
Vision metadata and inference telemetry from local hardware.
Model to IoT orchestration and NPU-based system telemetry.
Kinematics and spatial sensor telemetry for remote control.
Native support for multiple communication standards for cross-device compatibility.
Publish/subscribe protocol for IoT. Supports MQTT v3.1.1 and v5.0 features including topic aliases and shared subscriptions.
UDP-based variant for constrained devices and non-TCP networks like Zigbee or Bluetooth LE.
Full-duplex communication over single TCP connections for real-time browser-based telemetry.
Open registries that give agents, harnesses and orchestrators a declared source of truth instead of an assumption.
Agents are being given control of physical infrastructure. An agent that cannot look up a component will infer it, and an inferred pin or voltage is a fault with physical consequences.
These registries give agents, provisioning harnesses and orchestrators a declared source of truth: the commands a device accepts, the ranges its readings fall in, and what has been confirmed on hardware. A conforming tool should refuse anything outside it.
Certified hardware definitions. Wiring, drivers, units and operating ranges, verified against physical hardware.
Provisioning and operational capability, published by vendors and the community. New hardware support without vendor lock-step.
Certified runtimes and inference libraries per board. Toolchains resolved, not rebuilt.
Automated provisioning per platform. Device software generated and proven against live readings before handover.
Portable orchestration logic bound to device classes. One deployment pattern across heterogeneous estates.
Design time authoring, validated against registered capability. Invalid commands cannot be published.
Every entry records what has been confirmed on hardware, by whom, and what has not. Definitions are versioned and revisioned, so a change is always traceable.
Connect local LLMs to the IoT platform via the Model Context Protocol.
Define custom MCP tools and resources directly within the platform UI for specific hardware telemetry.
Bind specific MQTT topics to MCP tool-calls, allowing LLMs to read/write device states securely.
Local models execute these tools via NPU-accelerated telemetry, closing the loop between AI and IoT.
Real-time command routing triggers physical hardware actions based on LLM context and logic.
Automated IoT connected Digital Twins receive and display real-time data and machine learning predictions, allowing you to see faults and predicted faults in real-time.
Reference code and SDKs for Enterprise and Maker hardware integration with the IntelliConnect broker.
C++, MicroPython, and Rust drivers for microcontrollers including ESP32 and STM32.
Nodes to bridge local ROS 2 DDS telemetry to the cloud via MQTT v5 tunnels.
Boilerplates for streaming metadata from NVIDIA Jetson and OpenVINO inference engines.
Node.js and Go libraries for interacting with the platform management API.
Pre-registering for the IntelliConnect Beta provides early access to infrastructure features, the development roadmap, and early-adopter pricing tiers.
Access to the broker, MCPs, rule engine, and developer console during launch phase.
Direct communication channel with the engineering team for integration assistance.
Beta testers secure specialized pricing tiers for both Enterprise and Maker accounts.