AI Demos

Ultra-Low Power in Action

End-to-end demos run on Ambiq evaluation boards, trained and deployed with our ADKs, integrated with front-end sensors, and built using the neuralSPOT SDK. These demos demonstrate that advanced models can operate in real time on battery power—showcasing Apollo’s ultra-low-energy performance.

Where to Buy

Design Resources

Highlights

End-to-End, On-Device

Trained and deployed with Ambiq ADKs, wired to front-end sensors, and built with neuralSPOT firmware—real-time inference running entirely on the board.

Ultra-Low-Power

Advanced models on battery power. Each demo is tuned to showcase Apollo’s minimal latency, small footprint, and ultra-efficient energy use.

Open Source on GitHub

All demo code, ADK recipes, and neuralSPOT firmware are open-source—clone, fork, and customize. Contribute, report issues, and reuse in your product workflow.

Powerful Dashboards

Tileio brings live charts, interactive tiles, and device controls together—right in your browser. Connect via WebUSB/ WebBLE, tweak parameters, and visualize results in real time—no installs, no drivers.

Available Demos

Speech Enhancing Demo

NNSE is Ambiq’s AI speech denoiser, a real-time model that removes background noise (such as hair-dryer noise) from human speech.

PPG Codec Demo

Ambiq’s PPG Codec demo is a real-time AI model that compresses PPG data up to 16× while preserving signal quality, enabling longer battery life and continuous monitoring in smartwatches, smart rings, fitness bands, and remote patient devices.