Know your score
before the exam
knows it.
23 task types, scored the second you finish — with the trait-by-trait breakdown that names the mark you dropped and why.
4.8 from 42,800 learners in 38 countries
- Content
- 74/90
- Oral fluency
- 79/90
- Pronunciation
- 67/90
- Vocabulary
- 62/90
You lost 2 marks on content — the chart’s peak year and the units were never mentioned. Fluency and pronunciation were clean.
Every task type in the real exam — all 23
Practise smarter
One workspace for all four modules
Real exam timing, real marking, and feedback detailed enough to change what you do tomorrow.
Prompt
Look at the text below. In 40 seconds you must read this text aloud as naturally and clearly as possible.
Speaking — recorded in the browser, scored on fluency, pronunciation and content.
The loop
Four steps, then repeat
Every cycle moves your predicted score. Most learners run it three times a week.
Set your target
Tell us the score you need and your exam date. Every screen is measured against that number from then on.
Practise a task type
Answer under exam timing and get the score plus the trait breakdown before you move on.
Sit a mock test
A timed run with a section clock, question navigator and autosave, then a full report module by module.
Fix the weak spot
Analytics rank your task types worst to best. Work down that list until the prediction clears your target.
Why it works
Feedback specific enough to act on
Most platforms give you a number. This one names the trait that cost the mark, the word you skipped and the pair you put out of order.
Speaking, analysed in the browser
Record, play back, and see words per minute, hesitation count and every word you skipped — highlighted against the reference text, no upload and no waiting.
Your recording · Read Aloud
The committee agreed that renewable capacity should double within the decade, provided that storage costs keep falling.
Exam timing from the first click
Preparation timers, answer timers and section countdowns match the real thing, so test day holds no surprises.
Analytics that name the weak spot
A score trend, a module radar and accuracy by question type — so your study plan writes itself.
Everything on the 10–90 scale
Raw marks map through one published table, so a practice score means exactly what a mock score means.
Results
People who got the score they needed
“I had sat the test twice and kept landing on 71. The trait breakdown showed my writing form mark was zero every time because I was going over the word limit. Fixed that one habit and cleared 79.”
71 → 82 +11
“The speaking analysis was the thing. Seeing that I was averaging 96 words a minute with four long pauses explained a fluency score I had been arguing with for months.”
64 → 76 +12
“Re-order paragraphs used to be a coin flip for me. The adjacent-pair marking made it obvious that I was getting two out of four right and losing the rest to one bad guess at the opening.”
68 → 79 +11
“I used the analytics page as my study plan. Worst question type at the top, work down the list. Six weeks later the predicted score matched what I actually got, almost to the point.”
59 → 74 +15
“Write from dictation went from my worst task to my most reliable one, purely because the word-by-word highlighting showed me I was mishearing plurals rather than whole words.”
62 → 73 +11
“The mock test report is the closest thing to the real score report I have seen. Same shape, same modules, and it arrived instantly instead of five days later.”
66 → 78 +12
“I was practising for hours without improving because I never checked the answer key properly. Being forced to see the model answer after every single question changed how I studied.”
57 → 70 +13
“Timed practice from day one was uncomfortable and completely worth it. On the actual test day nothing about the pacing surprised me.”
54 → 68 +14
Questions
Everything else, answered
Reading and listening tasks are marked against the answer key using the published PTE rules - one mark per correct blank, negative marking on multiple-answer questions, one mark per correctly ordered adjacent pair, and so on. Writing is scored by a rule-based engine that measures form, spelling, mechanics, lexical variety and how much of the prompt you covered. Speaking is measured from your recording: words per minute, hesitation count and overlap with the reference text. Every raw score is then mapped onto the 10-90 scale through a single published table.
Not identical, and we label it clearly wherever it appears. The real exam uses a trained acoustic model that scores pronunciation at the phoneme level. Our analysis measures pace, pausing, speech continuity and word overlap with the reference text, which tracks fluency closely but is a proxy for pronunciation. Treat it as a reliable guide to your fluency and content, not a guarantee of the exact pronunciation sub-score.
PTE Academic is for university admission and most skilled migration routes. PTE Academic UKVI is the same test taken under UK Home Office secure conditions, required for certain UK visas. PTE Core is for Canadian economic immigration and tests everyday rather than academic English, replacing the essay with a workplace email task. Set your exam in your profile and the practice hub filters itself to the tasks you will actually sit.
Yes. Any built-in laptop microphone or headset works. Your browser will ask for permission the first time you open a speaking task, and the recording stays on your own machine until you submit it. On Chrome and Edge the browser also produces a live transcript, which unlocks the content and pace analysis.
The predicted score is the average of your module scores across your recent attempts, on the same 10-90 scale. It becomes meaningful once you have attempted a spread of task types - a prediction built on five reading questions tells you about reading, not about your overall score. The coverage figure on your dashboard shows how much of the bank you have sampled.
Yes. Institute plans cover multiple learners under one account, with a shared question bank, per-learner analytics and a teacher view that ranks the class by task type. Message us through the support page and we will set it up.
Still stuck? Ask us on the support page.
One scored question tells you
more than an hour of theory.
Free to start. No card. Your first trait breakdown lands in about ninety seconds.