Speech Recognition Engine

Quickly develop projects based on speech recognition, with a powerful and ready-to-use software library

The easy-to-use software library that leverages AI/ML to recognize voice commands in 40 languages – no vocal training or Internet connection needed

The Arduino Speech Recognition Engine offers the quickest and easiest way to start talking to and with machines. Its extensive software library was developed by worldwide speech recognition leader Cyberon with ease of use and compatibility in mind, so you can instantly integrate new applications – even in existing solutions – and start using your voice to interact with devices.

Speech Recognition Engine is compatible with multiple Arduino boards and the Arduino IDE, and requires no additional hardware, software or Internet connectivity. Its AI/ML engine understands commands defined through text input in 40+ languages – regardless of the speaker’s actual voice, tone or accent. This means you can quickly configure multiple wake-up words and sequences, and don’t have to retrain for different users.

Speech Recognition Engine is ideal for industrial and building automation applications such as hands-free control of machinery, equipment and human-machine interactions based on voice recognition: information kiosks, vending machines and lockers, smart beds in hospitals, emergency alert systems and more.

Speech Recognition Engine is able to listen to anyone speaking to it, ignoring any background noise.

Ready to test it?

Get started with a free trial license.


Download, integrate, done: using a software library makes speech recognition easy to integrate and use


Using text input to define voice commands translates into instant results and flexibility over different tones and accents


Add it to your existing or new projects, in combination with different elements in the Arduino ecosystem


Build your next smart device to understand every major language in the world: rhwydd! (that’s “easy”, in Welsh)

Key benefits
Key benefits
  • Powerful, yet extremely easy-to-use library with integrated AI/ML engine for phoneme-based modeling
  • Recognition of multiple wake-up words and sequences of commands
  • No vocal training required, commands configurable through text input
  • Support for 40+ languages, independent from accent variations
  • One configuration for multiple speakers, without retraining
  • Recognition on the edge, no need for additional HW/SW or connectivity
  • Suitable for noisy environments
  • Expand existing projects with speech recognition capabilities
  • Compatible with multiple Arduino Nano and Portenta products
  • Compatible with Arduino IDE and Arduino CLI
License specs
Speech Recognition Free Trial Speech Recognition license Pro License
n. of dataset* 1 1 Unlimited
n. of triggers 1 1 Unlimited
n. of commands 20 max 20 max Unlimited***
recog. times 50 Unlimited Unlimited
delay in trigger mode** 20s No No

*The dataset cannot be changed after deploying it and it is bound to a single Arduino board **Delay between entering the Trigger Mode and the recognition of the Wake-up Word ***Number of commands depend on available hardware resources

Need industrial-grade speech recognition?

Get your Professional License for advanced users