neuralSPOT SDK for Edge AI Development

Edge AI Made Easy

neuralSPOT™ is an open-source, real-time, OS-agnostic SDK for Ambiq’s Apollo SoCs, designed to simplify edge AI development on ultra-low-power devices. With integrated support for TFLM, heliaRT™, and heliaAOT™, neuralSPOT enables developers to build, optimize, and deploy high-performance AI applications faster across wearables, healthcare devices, smart home products, and industrial IoT systems.

Design Resources

Key Features of neuralSPOT SDK

01

Accessible Edge AI Tools

Open source, OS-agnostic, and real-time Al development toolkit built to streamline Al development with Ambiq’s range of ultra-low power SoCs.

02

Customizable AI Workflows

Everything the developer needs to develop, modify, train, and deploy sophisticated features on the Apollo family of SoCs.

03

AI Model Profiling and Optimization

Deploy, Fine Tune, and Profile AI models on all Ambiq development boards.

04

Rapid AI Prototyping

neuralSPOT streamlines AI features with simple app code, straightforward device I/O, and plug-and-play PC integration.

Bridging the Edge AI Divide

neuralSPOT SDK

neuralSPOT SDK

Edge AI Application Prototyping Library Simplifies the embedded development Frees AI Developers to focus on AI

neuralSPOT Autodeploy

neuralSPOT Autodeploy

Easy AI-to-firmware conversion Sophisticated Profiler and Validator Zero friction AI experimentation

Optimized AI Runtimes

Optimized AI Runtimes

neuralSPOT makes choosing a runtime easy by seamlessly integrating LiteRT, heliaRT and heliaAOT

AI Development Kits for neuralSPOT SDK

heartKIT

Detect heart statistics and conditions based on ECG sensor readings
Heart Rate, HRV, Afib detection
More added regularly

physioKIT

Optical-sensor based health statistics
SPO2, Heart Rate, Afib, cuff-less blood pressure monitoring

sleepKIT

Real-time: Inference is performed in real-time on battery-powered, edge devices
Efficient: Leverage modern AI techniques coupled with Ambiq’s ultra-low power SoCs
Extensible: Easily add new tasks, models, and datasets to the framework
Accurate: Achieve SoTA results with stringent resource constraints

Speech Enhancement

Eliminates background noise such as leaf-blowers, music, and restaurant noise from speech
Can be used to increase intelligibility and improve speech recognition accuracy

Speaker ID

Identifies speech belonging to a specific individual
Supports multiple individuals
Can be trained on new individuals on-device

AutoDeploy

Automatically analyzes, deploys, profiles, and packages AI models for Ambiq devices in a single workflow
Converts TFLite models into optimized embedded applications and static libraries
Profiles performance, memory usage, and power consumption directly on target hardware

Where to Buy

Apollo510 SoC 250Mhz 3.75MB BGA 2

Apollo510 SoC 250Mhz 3.75MB BGA

Apollo510 SoC, BGA

Apollo4 Blue Lite SoC 192 MHz 1.4MB 2

Apollo4 Blue Lite SoC 192 MHz 1.4MB

Apollo4 Blue Lite SoC

Apollo3 Blue Plus SoC 96 MHz 768KB BGA 2

Apollo3 Blue Plus SoC 96 MHz 768KB BGA

Apollo3 Blue Plus SoC

Apollo4 Blue Plus 192 MHz 2.75MB BGA

Apollo4 Blue Plus HexSPI 192MHz BLE 5.1

Apollo4 Blue Plus SoC

Apollo3 Blue Thin 96 MHz 384KB WLCSP 2

Apollo3 Blue Thin 96 MHz 384KB WLCSP

Apollo3 Blue Thin

Design Resources

Additional Documentation

Set favorites
Get update notifications
Use search filters

Visit Content Portal

Video Library

Frequently Asked Questions (FAQs)

neuralSPOT is Ambiq’s open-source AI software development kit designed for edge AI applications on the ultra-low power Ambiq SoCs.
neuralSPOT is optimized for Ambiq’s portfolio of ultra-low power SPOT-based SoCs with updates to support future silicon releases.
neuralSPOT simplifies AI model integration, optimization, testing, and deployment for embedded edge AI applications, reducing development complexity and accelerating time to deployment.
heliaRT provides an ultra-low power AI runtime that improves inference performance, memory optimization, and power efficiency for edge AI workloads — all tuned specifically for Ambiq silicon.
heliaAOT is an ahead-of-time compiler that converts LiteRT models into optimized standalone C inference modules for ultra-low power AI deployment on Ambiq SoCs.
Yes. neuralSPOT is designed for ultra-low power edge AI applications where battery efficiency is critical.
No. neuralSPOT is purpose-built for Ambiq’s portfolio of ultra-low power SoCs, leveraging the hardware architecture to deliver optimizations that general-purpose SDKs can’t match. From model integration and testing to deployment, every layer of neuralSPOT is tuned for Ambiq silicon — making Ambiq the best platform for edge AI development.