A method made of small steps
Your ears first, a little practice every day, and one fix at a time.
Six principles
The course follows six rules. The bot keeps to them, so you do not have to think about them.
- 1
Listen before you speak
Every lesson starts with your ears. It is hard to say a sound that you cannot hear yet.
- 2
A little every day
One new lesson opens each day. You can repeat any lesson that is already open, as often as you like.
- 3
Slow, then medium, then natural
You hear each line slowly first. The speed goes up during the week, when your mouth is ready for it.
- 4
One sound for a week
You stay with one small group of sounds for seven days. A new sound needs time to become a habit.
- 5
One fix at a time
After each recording you get one thing to try. You never get a long list of mistakes.
- 6
Show the progress, claim only what we measure
Each week you hear your day 1 next to your day 5. We tell you what we measured, and nothing more.
The week, day by day
Every unit takes one week and has the same seven days. The model voice gets faster as the week goes on.
- Day 1
Listen
You hear pairs of words and tap the one that was said. Then you hear the week's short story. You make one first recording, so we both know where you start.
Model voice: slow
- Day 2
Shape
You practise the week's sounds in pairs, inside short phrases. A picture shows the position of your mouth.
Model voice: slow
- Day 3
Phrase
You say phrases and short sentences. Listen, pause, repeat.
Model voice: medium
- Day 4
Shadow
You speak along with the week's short story, one line at a time.
Model voice: natural
- Day 5
Perform
You record the whole story with no model to follow. Then you hear it next to your recording from day 1.
No model voice
- Day 6
Review
You repeat the items that were hardest for you this week, and a few from earlier weeks.
Model voice: mixed
- Day 7
Rest
A rest day. Your weekly report arrives. There is a one-minute practice if you want it.
Every recording is a phrase or a sentence, never a single word. Single words are too easy for a machine to mishear.
Why short and daily
Pronunciation is a habit of the mouth. Your tongue and lips learn new movements, like in a sport.
We believe a little practice every day builds that habit better than one long lesson a week. So a lesson takes about 10 minutes, and one new lesson opens each day.
If you miss a day, your lesson waits for you. You go on from where you stopped.
Why listening comes first
It is hard to say a difference that you cannot hear yet.
So each week begins with ear training. You hear a word, for example "ship" or "sheep", and tap the one that was said. Only then do you record.
The model voice is AI-generated. It speaks General American English. You hear it slowly first, then at medium speed, then at natural speed.
Your before and after
On day 1 of each week you make a first recording of the week's short story. On day 5 you record it again, and you hear the two side by side.
At the start and at the end of the course you do the same with a longer text.
We score both recordings in the same way, so the comparison is fair.
What the bot can judge, and what it cannot
The bot checks whether two AI listeners understood each word. This fits sounds that turn one word into another, like the vowels in "ship" and "sheep".
The bot names a sound only when the exercise was made to test that sound. In other exercises it names the word, not the sound.
Word stress, rhythm and intonation are harder for software to judge. In those lessons you play your own recording and the model, and you compare them yourself.
We would rather show no score than a score we cannot stand behind.
Try the first week
The first unit is free. It is the fastest way to see if the method suits you.
First week free · no card