Developing AI applications for resource-constrained devices can be slow and complex, requiring developers to stitch together toolchains and spend time optimising models for power and latency. Ambiq’s latest release of neuralSPOT is designed to make that process easier.
The company has announced neuralSPOT SDK and Toolkit V1.2.0 Beta, now supporting the Apollo510 and Apollo510B SoCs. These are Ambiq’s highest performance, most power-efficient devices to date, making them well-suited for advanced edge AI tasks such as speech processing, health monitoring, and activity recognition.
New Features for Apollo510 Developers
Version 1.2.0 unlocks full Apollo510 performance profiling with new AI characterisation tools, helping engineers measure and tune inference speed and power draw. The release introduces HeliaRT Beta, Ambiq’s Apollo-specific AI runtime, which the company says can deliver up to three times faster inference and improved energy efficiency compared to LiteRT-based implementations. An experimental HeliaAOT compiler is also included for ahead-of-time compilation and deployment, targeting performance-critical use cases.
Rich Example Library
The SDK comes with more than a dozen ready-to-use model examples, covering applications such as human activity recognition, ECG monitoring, keyword spotting, speech enhancement, and speaker identification. These implementations give developers a fast way to move from concept to running prototypes.
Building on Apollo510 Hardware
The Apollo510 SoC family has become a flagship platform for Ambiq, winning the Best Hardware Award at Embedded World 2024 for its combination of high performance and ultra-low power consumption. neuralSPOT integrates with Ambiq’s heartKIT™ AI SDK, which won the Best AI Solutions Award at Embedded World 2025, giving developers access to pre-built heart vital sign analytics models.
For developers working on edge AI projects, neuralSPOT V1.2.0 Beta offers a complete toolkit to speed development, benchmark performance, and deploy optimised AI models on Apollo510-based devices.
Learn more and read the original article on www.ambiq.com
You may also like
