HeliaRT

Ultra-Efficient Runtime

HeliaRT is an ultra-efficient AI runtime built on TensorFlow Lite for Microcontrollers, optimized for Ambiq’s Apollo family of ultra-low-power SoCs. It simplifies development and accelerates deployment of high-performance, energy-efficient AI applications at the edge.

Video Library

Design Resources

HeliaRT highlights

Optimized for Performance

Utilizes Apollo SoC’s M-Profile Vector Extensions (MVE) and DSP capabilities to accelerate AI computations and maximize efficiency.

Custom AI Kernels

Includes purpose-built kernels optimized for Apollo510’s vector acceleration, unlocking enhanced speed and performance for edge applications.

High Performance, Ultra-Low Power

Enables developers to create responsive AI applications with minimal energy usage—ideal for battery-powered and always-on devices.

Seamless Compatibility

Fully backwards-compatible with TensorFlow Lite for Microcontrollers and supports the entire Ambiq SoC lineup, ensuring flexibility across diverse hardware platforms.

Turbocharged Inference

HeliaRT delivers up to 7× faster inference than TensorFlow Lite for Micro, while offering extensive kernel support across AI layer types—no fallbacks, zero performance loss

Design Resources

HeliaRT Developer Hub

Explore step-by-step guides, hands-on examples, and performance benchmarks to kickstart your HeliaRT integration

Ambiq Al Github

Dive into neuralSPOT AI SDK & toolkit and our heartKIT, sleepKIT, and soundKIT ADKs—everything you need in one place to build, deploy, and optimize AI on Ambiq devices