Our method and your AI teacher

Start free

The method

PracticeAloud is built around a simple idea: a language is a speaking skill, not a stack of completed flashcards. So every session centers on speaking aloud, and everything else — the plan, the homework, the report cards — exists to make that conversation better.

• A CEFR curriculum. The course skeletons for every level are written and vetted by a credentialed teacher-methodologist, not generated on the fly.
• One AI teacher. You have a consistent AI teacher who remembers every session. Inside scenarios they play different characters, but the thread of your learning never breaks.
• A closed loop. Conversation → analysis → plan → homework → next conversation. Mistakes from a lesson become exercises; exercises become material for the next lesson.
• Human quality review. A real teacher spot-checks lessons and reports and adjusts study plans where needed. Automation never silently overrides a human decision.

Speaking first, and why

A great deal of English study is silent: reading, apps, grammar drills, films. It builds recognition — you understand far more than you can say — and recognition does not turn into speech on its own. Producing a sentence under time pressure, with someone waiting for the answer, is a separate skill, and it grows only when it is used.

That is why the lesson is a conversation and not an exercise with a voice button. The teacher asks, you answer, the answer is imperfect, and the imperfection is the material. Everything in the method is arranged so that you spend the lesson talking and the rest of the week fixing what the talking revealed.

A curriculum, not an improvisation

Behind every plan is a course skeleton for your level, from A1 to C1: the objectives a learner in that band needs, the order they are best taken in, and what it means for one to be done. A credentialed teacher-methodologist writes and vets those skeletons. The AI personalises within them — it picks scenarios that fit your goal, paces the plan to the time you actually have, and brings back what you got wrong — but it does not invent the course.

The point is consistency. Two learners at the same level get the same pedagogy, not two different improvisations, and a study plan reads like a syllabus rather than a random list of topics. You can open yours and see what is coming and why.

What “mastered” means here

Every objective in your plan has a state: new, practicing, or mastered. It moves on evidence from your lessons and homework, and it moves slowly on purpose — one good day is not mastery, so an objective stays in practicing until it has held up more than once. Things you have mastered still come back for review, on a widening schedule, because that is how they stay mastered.

This is also what your homework is built from. An exercise exists because a specific objective, or a specific mistake, is due — not because a content library had something at your level.

One teacher, and what they remember

Your teacher keeps a record: your level and how confident the estimate is, your goal and timeline, your interests, the mistakes you keep making, and how your homework went. Every lesson begins from that record and from a summary of the last one — the teacher can pick up a thread from a week ago, or point out that a pattern you struggled with has stopped appearing.

In a scenario the teacher plays a character: an interviewer, a doctor, a landlord. The character changes; the teacher does not. Why a tutor’s memory matters is a longer argument; the short version is that a tutor who forgets you can only ever teach the first lesson.

Corrections and your native language

Some corrections happen in the conversation, when a small fix keeps you talking. All of them are recorded, and the ones that matter come back — in your report card, with your own sentence in it, and as homework. Correcting everything in the moment would turn a lesson into a dictation, and a mistake you fix yourself the next day sticks better than one you were interrupted over.

When something is unclear, the teacher explains it in your native language, out loud, and returns to English. How much of that you hear depends on your level: a beginner gets a lot of it, and it fades as you no longer need it.

Human quality review

Behind every PracticeAloud plan is a human method, not a model left to improvise. A qualified teacher-methodologist writes and vets the CEFR course skeletons for each level, defines the rules the AI teacher follows when correcting mistakes and running a lesson, and spot-checks real lessons, report cards, and study plans to keep the quality high. Automation never silently overrides a human decision.

That works in both directions. A teacher can edit a learner’s study plan by hand, and when the automated analysis later wants to restructure that plan, it proposes a new version for a person to approve rather than overwriting what the teacher decided.

That is what keeps your study plan grounded in a real teaching method — not a random list of topics.

Honest reports

A report card says what went well, lists the corrections with your own sentences in them, records the words you picked up, and counts the minutes you actually spoke. It does not say your level grew by a percentage, because nobody can measure that from one lesson. Once a week your teacher adds a short note on where you are and what comes next. Progress in this method is something you can read, not a number that moves.

See it in practice

How it works walks through the intake, a lesson and the homework step by step. What a closed learning loop actually is explains why the loop, not the lesson, is the unit that matters. The first lesson is free.