Pronunciation Improvement

Make your speech clear with new feedback technology:


Learn which sounds need improvement and when you use intonation differently.


Practice from anywhere and anytime.


Keep track of your progress up to the smallest details.


No payments or credit card infromation is required.

Learn about our course


Our experts believe that most accents can be described with three parameters:

Durations of sounds

Vocal intonation

Articulation quality

How to make your pronunciation clear? Our answer is to become aware of your individual assumptions and tendencies in pronunciation and to have enough practice so that you use the right pronunciation unconsciously. Accent Hero uses modern speech recognition tools to provide you with feedback in real time, showing tips and comparing your pronunciation to the pronunciation of a native U.S. English speaker.

Accent Hero presents results visually, which helps to see a difference even if you don’t have an ear for music. Moreover, we’ll show you a reference transcription, just in case, if you’d wonder what a dictionary says is the right way.


The visualisation

Why one can't travel to an English-speaking country and by living in a culture pick the right pronunciation? Well, it turns out that most of us are much better at analysing visual information than minuscule differences in sounds. In addition, many people just don't have an ear for music.

To overcome the awkward analysis of sounds, Accent Hero, in addition to recordings of a native speaker, visualizes all the key differences in pronunciation, making it easy to see them.


Accent Hero uses two types of assessments to supplement your learning: assessments by people and feedback from the technology. Together they create a hybrid system with benefits of each of them and the drawbacks of none.









Technical notes

Accent Hero requires a desktop web browser that supports HTML5 Audio Capture API, which is supported by recent versions of all popular browsers - Chrome, Firefox, Opera, and Edge. Also, your browser will ask you to allow access to your microphone. Please, do so.

The accuracy of acoustic analysis depends on the level of noise in your surrounding environment, so if you get inconsistent results, please try practicing using a microphone headset and in a quieter room. Also, our algorithms are tuned to better detect pronunciation mistakes of non-native speakers of English.

The service uses statistical models to find and distinguish phonemes, guess stressed syllables, etc. Our model is based on data from more than a 1000 volunteers. Use of statistics means the more data we have the more accurate results will be. We are just starting, so if something doesn't work as you expected, please give the service a benefit of doubt.