DEEPCRAFT™ AI Suite

Your fast-track to intelligent, power-optimized Edge AI products

nobutton

About

DEEPCRAFT™ is Infineon's end-to-end Edge AI software suite for low-power, production-ready products 

DEEPCRAFT™ helps you take AI-enabled products to production. No matter where you are in your Edge AI journey, you can leverage our AI Suite to add low-power, intelligent features to your embedded devices. 

  • Build your custom models in DEEPCRAFT™ Studio, our development platform
  • Convert proprietary or open-source models from tflite, h5, keras and PyTorch with DEEPCRAFT™ Model Converter
  • Deploy production-ready DEEPCRAFT™ Solutions developed by Infineon's ML experts, with voice, audio and event detection models

Use the DEEPCRAFT™ AI Hub to find your easiest path to Edge AI enabled devices. 

Explore the DEEPCRAFT™ AI Hub
Download DEEPCRAFT™ Studio

DEEPCRAFT™ AI Hub is Infineon's centralized catalog of Edge AI models, tools, and solutions

The DEEPCRAFT™ AI Hub is a one-stop resource for Edge AI on Infineon MCUs. It hosts open-source models, pre-configured DEEPCRAFT™ Studio Accelerators (with the datasets, preprocessing, model architecture, and instructions you need to develop them into production-ready models), and DEEPCRAFT™ Solutions — each validated to fit within the power and memory constraints of PSOC™ Edge and other Infineon MCU families. You'll also find the platform and tools you need.

Use the Hub to:

  • Benchmark what's feasible on Infineon MCUs before committing engineering resources
  • Download pre-trained models and Studio Accelerators
  • Review reference designs for smart home, wearables, home appliances, automotive, industrial, and more
Visit the DEEPCRAFT™ AI Hub

DEEPCRAFT™ Studio is an end-to-end platform for building production-grade custom Edge AI models on Infineon MCUs

DEEPCRAFT™ Studio is Infineon's end-to-end development platform for custom Edge AI. It covers data collection, labeling, preprocessing, cloud-based training, evaluation, and hardware-optimized code generation — compressing months of engineering into a guided workflow. Built on over a decade of Edge AI customer-project experience, the platform is data-agnostic and supports time series input data (such as audio, radar, vibration, motion, etc.) as well as vision input data.
 

Key features and capabilities include:

  • Graph UX interface — no-code model building with live evaluation
  • DEEPCRAFT™ Studio Accelerators: pre-configured projects with datasets, preprocessing, and model architectures for many use cases
  • Cloud training; models downloaded as TensorFlow .h5, with simple conversion to C-code for deployment on AURIX™, TRAVEO™, XMC™, etc.
  • Direct code generation for PSOC™ Edge and PSOC™ 6
  • Transparent data policy designed to protect our customers' IP
  • Unlike generic ML frameworks or cloud-inference approaches, every component is tuned for power- and memory-constrained MCUs
Process Brand Circle
Process Brand Circle
Process Brand Circle
Learn more about DEEPCRAFT™ Studio
Download DEEPCRAFT™ Studio

DEEPCRAFT™ Model Converter turns PyTorch and TensorFlow models into hardware-optimized code for Infineon MCUs

DEEPCRAFT™ Model Converter bridges popular AI frameworks and Infineon's embedded hardware. It ingests PyTorch (.pt2) and TensorFlow (.h5 /keras/.tflite) models and generates deployment-ready C code for PSOC™ Edge and PSOC™ 6 — including optional 8-bit network quantization, and sparsity-based memory optimization to fit models within MCU budgets.

Use DEEPCRAFT™ Model Converter when you:

  • Already have a proven model from an existing AI pipeline and need to run it on a power-constrained MCU
  • Want framework portability instead of lock-in to a proprietary, single-vendor toolchain
Learn more about DEEPCRAFT™ Model Converter
DEEPCRAFT™ AI Suite offers production-ready Edge AI solutions built by Infineon ML experts for fast integration into your products

DEEPCRAFT™ Solutions eliminate the need for in-house model development. Each Solution comes pre-trained, extensively tested on Infineon hardware, and tuned for low-power operation. Use them to ship intelligent features without adding ML engineering headcount.

 

DEEPCRAFT™ Solutions include:

  • DEEPCRAFT™ Voice Assistant — Speech-to-Intent with <1 mW always-on wake-word detection; 7 mW for full assistant with 20 commands.Designed for PSOC™ Edge
  • DEEPCRAFT™ Audio Enhancement — Noise Suppression, Acoustic Echo Cancellation, Audio Scene Analysis, and multi-mic Beamforming. Designed for PSOC™ Edge
  • DEEPCRAFT™ Ready Models — We currently have eight Ready Models, with use cases in applications such as smart home, wearables, home appliances, automotive, industrial, etc. These models are designed to run on MCUs, with as low as 3kB RAM and 15kB Flash.
DEEPCRAFT™ Ready Models
DEEPCRAFT™ Voice Assistant
DEEPCRAFT™ Audio Enhancement

DEEPCRAFT™ is Infineon's end-to-end Edge AI software suite for low-power, production-ready products 

DEEPCRAFT™ helps you take AI-enabled products to production. No matter where you are in your Edge AI journey, you can leverage our AI Suite to add low-power, intelligent features to your embedded devices. 

  • Build your custom models in DEEPCRAFT™ Studio, our development platform
  • Convert proprietary or open-source models from tflite, h5, keras and PyTorch with DEEPCRAFT™ Model Converter
  • Deploy production-ready DEEPCRAFT™ Solutions developed by Infineon's ML experts, with voice, audio and event detection models

Use the DEEPCRAFT™ AI Hub to find your easiest path to Edge AI enabled devices. 

Explore the DEEPCRAFT™ AI Hub
Download DEEPCRAFT™ Studio

DEEPCRAFT™ AI Hub is Infineon's centralized catalog of Edge AI models, tools, and solutions

The DEEPCRAFT™ AI Hub is a one-stop resource for Edge AI on Infineon MCUs. It hosts open-source models, pre-configured DEEPCRAFT™ Studio Accelerators (with the datasets, preprocessing, model architecture, and instructions you need to develop them into production-ready models), and DEEPCRAFT™ Solutions — each validated to fit within the power and memory constraints of PSOC™ Edge and other Infineon MCU families. You'll also find the platform and tools you need.

Use the Hub to:

  • Benchmark what's feasible on Infineon MCUs before committing engineering resources
  • Download pre-trained models and Studio Accelerators
  • Review reference designs for smart home, wearables, home appliances, automotive, industrial, and more
Visit the DEEPCRAFT™ AI Hub

DEEPCRAFT™ Studio is an end-to-end platform for building production-grade custom Edge AI models on Infineon MCUs

DEEPCRAFT™ Studio is Infineon's end-to-end development platform for custom Edge AI. It covers data collection, labeling, preprocessing, cloud-based training, evaluation, and hardware-optimized code generation — compressing months of engineering into a guided workflow. Built on over a decade of Edge AI customer-project experience, the platform is data-agnostic and supports time series input data (such as audio, radar, vibration, motion, etc.) as well as vision input data.
 

Key features and capabilities include:

  • Graph UX interface — no-code model building with live evaluation
  • DEEPCRAFT™ Studio Accelerators: pre-configured projects with datasets, preprocessing, and model architectures for many use cases
  • Cloud training; models downloaded as TensorFlow .h5, with simple conversion to C-code for deployment on AURIX™, TRAVEO™, XMC™, etc.
  • Direct code generation for PSOC™ Edge and PSOC™ 6
  • Transparent data policy designed to protect our customers' IP
  • Unlike generic ML frameworks or cloud-inference approaches, every component is tuned for power- and memory-constrained MCUs
Process Brand Circle
Process Brand Circle
Process Brand Circle
Learn more about DEEPCRAFT™ Studio
Download DEEPCRAFT™ Studio

DEEPCRAFT™ Model Converter turns PyTorch and TensorFlow models into hardware-optimized code for Infineon MCUs

DEEPCRAFT™ Model Converter bridges popular AI frameworks and Infineon's embedded hardware. It ingests PyTorch (.pt2) and TensorFlow (.h5 /keras/.tflite) models and generates deployment-ready C code for PSOC™ Edge and PSOC™ 6 — including optional 8-bit network quantization, and sparsity-based memory optimization to fit models within MCU budgets.

Use DEEPCRAFT™ Model Converter when you:

  • Already have a proven model from an existing AI pipeline and need to run it on a power-constrained MCU
  • Want framework portability instead of lock-in to a proprietary, single-vendor toolchain
Learn more about DEEPCRAFT™ Model Converter
DEEPCRAFT™ AI Suite offers production-ready Edge AI solutions built by Infineon ML experts for fast integration into your products

DEEPCRAFT™ Solutions eliminate the need for in-house model development. Each Solution comes pre-trained, extensively tested on Infineon hardware, and tuned for low-power operation. Use them to ship intelligent features without adding ML engineering headcount.

 

DEEPCRAFT™ Solutions include:

  • DEEPCRAFT™ Voice Assistant — Speech-to-Intent with <1 mW always-on wake-word detection; 7 mW for full assistant with 20 commands.Designed for PSOC™ Edge
  • DEEPCRAFT™ Audio Enhancement — Noise Suppression, Acoustic Echo Cancellation, Audio Scene Analysis, and multi-mic Beamforming. Designed for PSOC™ Edge
  • DEEPCRAFT™ Ready Models — We currently have eight Ready Models, with use cases in applications such as smart home, wearables, home appliances, automotive, industrial, etc. These models are designed to run on MCUs, with as low as 3kB RAM and 15kB Flash.
DEEPCRAFT™ Ready Models
DEEPCRAFT™ Voice Assistant
DEEPCRAFT™ Audio Enhancement
Documents

Design resources