Sit all four modules. Free. Marked on all four.
Most free practice sites give you Listening and Reading, because those mark themselves. Writing and Speaking need an examiner — so they are usually paid, or simply missing. This site marks all four against the official assessment criteria and shows you the reasoning behind every band. That is the part that was missing when this was built, and it is why it exists.
Practice tests in an interface built to match the real computer-delivered exam — then a band score against the four official assessment criteria, with the evidence behind every judgement.
The full test under exam conditions, or one module on its own. Timed, audio plays once, band score at the end with the evidence behind it.
Start a test →Drill a single question type — Listening maps, True/False/Not Given, matching headings, pie charts. No clock, audio replays as often as you like, every answer explained as you go.
Choose a drill →A test tells you where you stand. A drill is how you move. If your score has stopped improving across several papers, more papers are not the fix — that is what the right-hand door is for.
How the marking worksSit all four in exam order with a break before Speaking and get an overall band — or pick a single module below whenever you want to drill one skill. They are listed in the order the real test runs them.
4 recordings, 40 questions, 30 minutes. Audio plays once, as in the real test — then replay the exact moment each answer was spoken.
Start test →Timed passages, auto-marked instantly — then see exactly where in the passage every answer you missed was hiding.
Start test →Task 1 and Task 2, 60 minutes. Marked on Task Response, Coherence & Cohesion, Lexical Resource, and Grammatical Range & Accuracy.
Start test →All three parts with a real examiner, 11–14 minutes, recorded in your browser — then a band on all four speaking criteria.
Start test →All four modules are available, with 10 full papers each for Listening and Reading. Writing and Speaking were built first, because they are the two you cannot mark yourself — and the two where students otherwise have to rely on a teacher of unknown quality.
No email, no password. You get a short ID that stores your results so you can track progress across every module. Save it — it cannot be recovered.
Most automated IELTS checkers ask an AI “what band is this essay?” and report whatever comes back. That produces confident, inconsistent guesses. This works differently:
| Step | What happens |
|---|---|
| 1. Measure | Your response is analysed for length, paragraphing, sentence variety, cohesive-device use and overuse, lexical diversity and repetition. No AI — just counting, so it is identical every time. |
| 2. Score each criterion separately | Four independent judgements against the official band descriptors. A weak vocabulary score cannot drag down your grammar score. |
| 3. Compare to marked scripts | Every judgement is made against real scripts carrying official examiner bands, so it is a comparison rather than a guess. |
| 4. Quote the evidence | The marker must quote your own words before it may award a band. |
| 5. Score three times | Each criterion is marked three times and the middle score taken, so resubmitting the same response gives the same band. |
Treat any band as accurate to about ±0.5. That is roughly the level of agreement between two human examiners marking the same script — IELTS reports an inter-rater reliability of 0.92 for Writing, and even machine-marked Listening carries a standard error of about 0.43 bands. No marker, human or automated, produces a single exact truth.
This is practice, not a prediction. Use it to find weaknesses and track improvement, not to decide whether to book your test.