Less Data. More Intelligence.

Ambiq's compressionKIT is an advanced AI-powered codec that aims to lower the power and memory usage when handling continuous sensor data, such as from wearables and other edge devices, before it is stored, transmitted, or analyzed—enabling more intelligence with less data.

Design Resources

Video Library

compressionKIT Highlights

Press Release

Up to 20x Data Compression*

Achieves as much as 20x compression of continuous sensor data while preserving critical features for AI and analytics.

Up to 16x Lower On-device Memory Usage**

Enables longer data retention and reduces storage requirements.

Energy Multiplier

compressionKIT reduces transmission energy by sending fewer bits over the air, helping extend overall battery life.

Configurable Targets and Workflows

Supports adaptable edge-to-cloud processing, including real-time on-device inference, compressed cloud transfer, and customizable compression levels (2x –16x) for developers to balance performance and constraints.

Applications

Smart Watches

All-in-one wearables that track activity, heart rate, sleep, and notifications, while enabling advanced health insights like ECG and stress monitoring in a convenient, everyday device.

Smart Rings

Compact, discreet wearables designed for continuous health tracking, such as sleep quality, heart rate variability, and recovery without the bulk of a wrist-based device.

Fitness Bands

Lightweight and affordable trackers focused on core metrics like steps, calories, heart rate, and workouts to more advanced monitoring such as stress, sleep, and arrhythmia detection.

Remote Patient Monitors

Medical-grade devices used to track vital signs such as blood pressure, glucose, or cardiac activity, enabling clinicians to monitor patients in real time outside traditional healthcare settings.

Design Resources

Additional Documentation

Set favorites
Get update notifications
Use search filters

Access Content Portal

Design Resources

Baseline assumes uncompressed transmission of raw data. Results were measured on the Apollo510 using two-channel PPG sampled at 64 Hz in 4-second windows. Reported compression combines a fixed 16x reduction from the compressionKIT model with a dynamic entropy encoder, achieving average compression ratios of up to 20x over one minute of data. The data compression ratio at 20x is not deterministic.
Memory usage reduction is based on internal benchmarking on the Apollo510 of compressed versus uncompressed data storage under the same operating conditions. Actual savings may vary depending on signal characteristics, sampling rates, and application requirements.

Less Data. More Intelligence.

compressionKIT Highlights

Up to 20x Data Compression*

Up to 16x Lower On-device Memory Usage**

Energy Multiplier

Configurable Targets and Workflows

Applications

Smart Watches

Smart Rings

Fitness Bands

Remote Patient Monitors

Design Resources

Additional Documentation

Instant Download

Design Resources

*Up to 20x Data Compression

**Up to 16x Lower On-device Memory Usage

Video title