AI that helps youace the exam and land the admit.
Full-format mocks and sectional drills for English proficiency, admissions, Japan, Korea, Germany, Sweden and Finland.
Free plan included · No credit card · 14 exams supported
13 study destinations across 4 continents — practice, score and apply in one workspace.
The platform · eight tools, one workspace
Everything that turns a mock test into an admit.
Each panel below is a real screen from inside PrepGPT — so you can see exactly what every tool does, before you ever sign up.
Free to startPractice, scoring and interviews are free — Pro adds scholarship matching & professor outreach.
Mock tests
Sit the real exam, on the clock.
Q17. The library closes at
AI grading
Essays and speaking scored on the official band descriptors — with the exact sentence to fix.
7.5
Overall band · target 7.0
+0.5 vs last attemptFix first
There is many reason There are many reasons — to lift Grammar to 7.0.
University finder
Your grades, scores and budget become a ranked shortlist — reach, target, safety, each with a live match %.
Carnegie Mellon
M.S. Computer Science
University of Toronto
M.Sc. Computer Science
University of Auckland
M.Sc. Data Science
24 programs matched across 13 destinations
Scholarship finder
Funding you actually qualify for — ranked by fit and deadline.
Melbourne Research Grant
U of Melbourne · merit
Vanier Canada Graduate
U of Toronto · research
Fulbright Foreign Student
Carnegie Mellon · need
Visa interview AI
Rehearse visa and university admission interviews out loud — get a verdict, red flags and pronunciation after every answer.
Professor emails
Find the right professor, verify the email, and let AI draft personalized cold outreach.
Dr. S. Lang · Robotics Institute
s.lang@cs.cmu.edu
Your 2025 SLAM paper — prospective M.Sc. applicant
“Your group's loop-closure work is why I…”
Guided learning
Hints that teach — every wrong answer explained, then re-queued.
In line 12, “arrest” most nearly means…
Why B tempts you
“Arrest” usually means capture — but with “growth” it means to halt.
Adaptive study roadmap
A weekly plan built from your exam date and your weakest skills.
Map-labelling drill · 15 min
doneReading · Matching headings ×2
Targets weak: Reading 5.5Speaking Part 2 · full mock
NextSee it in action
Watch one mock test become a whole plan.
A short tour of the loop — score, fixes, roadmap, shortlist and interview prep, all in one workspace.
One engine · every exam
Every exam, on the real format.
Full mocks + sectional drillsHow it works
One loop. Three moves.
Every attempt turns into feedback, every piece of feedback turns into a sharper plan.
Take a real mock
Sit the test in exam mode: timed sections, locked audio, word limits, and the same pressure you will feel on test day.
Get graded, then coached
PrepGPT scores against the rubric, explains the miss, and shows the exact habit to fix before your next attempt.
Follow the updated plan
Your roadmap, admit chances, university shortlist, and scholarship matches refresh as your performance improves.
The loop keeps compounding
Mock results update lessons, applications, scholarships, and interview prep.
Exam engine
Practice that feels exactly like test day.
Timed sections, real question types, audio that behaves like the real thing — then a report that shows exactly where the points went.
- IELTS, TOEFL, PTE, Duolingo, SAT, GRE, GMAT, EJU, TOPIK, TestDaF, DSH, Tisus & YKI — full-length and sectional
- AI-generated question banks reviewed before they reach students
- Mistake heatmaps that learn your weak spots over time
The migration patterns of Arctic terns have long puzzled ornithologists. Recent satellite tracking reveals that these birds…
14. The writer suggests the tracking data was
Speaking · Part 2
Describe a place you remember well
Phoneme-level pronunciation analysis · Azure speech + Claude content scoring
86
Fluency
91
Pronunciation
84
Vocabulary
Speaking & writing AI scoring
Examiner-grade feedback, down to the syllable.
Two scoring engines, every exam. Speaking analyzed twice and writing graded on the official band descriptors — so you know whether to fix what you said or how you said it.
- Score speaking on pronunciation, fluency and prosody — word by word
- Check writing for grammar, spelling and lexical range
- Per-criterion bands on the official descriptors, never one vague number
- Every attempt and fix saved to your profile, on any device
Guided learning
It doesn't just score you. It teaches you.
Guided mode turns the AI into a tutor — hints before you answer, teaching after. Hearts and streaks keep the habit alive.
- Hints that nudge without spoiling — up to three per question
- Every distractor explained: know exactly why the trap worked
- FSRS spaced repetition schedules reviews at the forgetting edge
- Hearts, streaks and daily quests — practice becomes a habit
The word “novel” in line 12 is closest in meaning to…
7.0 → 7.5AI coach
“Novel” as a noun means fiction — but here it's an adjective. The context (“a novel approach”) signals new.
Your path · Reading
14-day streakSkimming & scanning
Map labelling
64%True / False / Not Given
Admissions OS
Your scores become a strategy.
When your scores move, your shortlist moves with them — programs, scholarships, SOPs and deadlines, all on one board.
Universities
Scholarships
SOP builder
Deadlines
Universities & scholarships across 13 destinations
Carnegie Mellon
M.S. Computer Science
University of Toronto
M.Sc. Computer Science
University of Melbourne
M.Sc. Computing
University of Auckland
M.Sc. Data Science
Writing band 7.0 → 7.5 would move Carnegie Mellon from reach to target. Want the drill plan?
Interview coach
Rehearse the conversation before it counts.
A strict AI officer runs the interview — real questions, real pressure. Then you get scores and every red flag a real officer would have noted.
86
Pronunciation
88
Fluency
95
Confidence
91
Content
F-1 Visa Officer
Live · strictness 4 of 5
Pronunciation · scored by AI
86/100Theirrobotics72labpublishedthepapermyundergraduatethesis61wasbuilton
ImproveStress the first syllable of “robotics” (ro-BOT-ics) and soften the “th” in “thesis” — drill these two and you clear 90.
Coach note
Strong specificity. Now add one sentence on year-two funding — financial ambiguity is the most common F-1 red flag.
6
exams, one engine
One interface for every major test path.
<60s
essay grading time
Rubric feedback while the attempt is still fresh.
7
study destinations
Each one supported end to end — admissions to visa prep.
$72K
avg funding tracked
Eligible awards surfaced before deadlines vanish.
Student outcomes
Quiet tool. Loud results.
Real results students shared after preparing with PrepGPT.
The writing feedback is the closest thing to a real examiner I’ve found. It told me exactly which paragraph was costing me the band.
I practiced the visa interview eleven times. The real one felt like a rerun — I knew every question before the officer finished asking.
The scholarship radar found a graduate award I had never heard of. That one match paid for my entire first year.
Instant, accurate evaluation — built to hit your target score.
Our AI scoring for English-language mock tests gives you instant, accurate performance evaluation on the official rubric — so every attempt targets exactly the score you need.
Scored on the real rubric
Writing graded against the official band descriptors — never one vague number.
Reasoning you can audit
Every criterion shows why, with the exact sentence to fix next.
Instant, every attempt
Full feedback in seconds, while the answer is still fresh in your head.
FAQ
Find answers to your common questions.
Everything you need to know before you start — accuracy, pricing, exams and more.
Is PrepGPT’s AI scoring accurate?
Yes. Writing is scored against the official band descriptors and speaking adds phoneme-level pronunciation analysis. Every score comes with per-criterion reasoning you can read and verify yourself — never one unexplained number.
How quickly are mock test results available?
In seconds. Full sectional feedback and band scores land while the attempt is still fresh, so you can fix the exact issue before your next try.
Will PrepGPT help me reach my target score?
That’s the goal. Set a target and every mock shows your gap, the “fix this first” item per section, and an adaptive roadmap that re-plans around your weakest skills as you improve.
How many mock tests can I take?
Practice is free to start — take mocks and sectional drills on the free plan. Pro removes limits and adds the admissions tools (scholarship matching and professor outreach).
Is PrepGPT free to start?
Yes — start free, no credit card. Mock tests, AI grading, guided learning, the visa & admission interview practice, the study roadmap and the university finder are all on the free plan.
Which features need Pro?
Pro unlocks the discovery and outreach tools — the scholarship finder and AI-drafted professor emails — plus unlimited practice. Everything else is free.
Which exams are supported?
Fourteen: IELTS Academic & General, TOEFL iBT, PTE Academic and the Duolingo English Test for English proficiency; SAT, GRE and GMAT for admissions; plus EJU (Japan), TOPIK (Korea), TestDaF and DSH (Germany), Tisus (Sweden) and YKI Finnish (Finland). Each runs as a dedicated, full-length flow, and new exams ship as data — so the list grows fast.
Can I practise the visa and university interview?
Yes. The interview AI rehearses both visa and university admission interviews out loud, then returns a verdict, the red flags an officer would note, and pronunciation feedback after every answer.
Does PrepGPT find universities and scholarships for me?
Your scores and budget become a ranked university shortlist — reach, target and safety, each with a live match %. The Pro scholarship finder then surfaces funding you actually qualify for, ranked by fit and deadline.
Can I start without a study-abroad plan?
Yes. Pick an exam and start practising in under a minute. The admissions tools activate whenever you’re ready — your scores carry over automatically.
Your admit letter is a plan, not a lottery.
Start with one mock test. PrepGPT builds the rest of the path with you.