How assessment works

Understand your strengths and what to practise next. Get task-specific AI feedback with criteria and examples.

AI practice feedback · Parliva scale 0–4 · Not an official exam score.

Feedback built around your task

For writing and speaking tasks, Parliva's AI checks your answer against the task you opened: its requirements, such as word count, and a set of criteria such as task completion, organisation, vocabulary and grammar.

For each criterion you get a band, a short example quoted from your own answer (for example « je suis allé »), and clear next steps.

Parliva's own 0‑4 practice scale

Bands run from 0 to 4: 0 Not demonstrated · 1 Limited · 2 Developing · 3 Effective · 4 Strong. This scale and its descriptions are written by Parliva for practice.

They are informed by the task descriptions that exam organisers publish. Parliva does not use the official, private grading grids, and the bands are not official scores, examiner-validated predictions or a certification. Parliva is independent and is not affiliated with or endorsed by the exam organisers.

Where the task information comes from

Task formats and requirements, such as task types and word counts, follow the public descriptions from the exam organisers:

The criterion descriptions and bands are Parliva's own practice descriptors, not text from these organisations. Always check the official pages for current exam rules.

Speaking: assessed from your transcript

In Voice speaking tasks, what you said is first turned into text. You can review that transcript and fix words the app misheard. The assessment then uses your reviewed transcript, and your result is labelled as based on the original or the edited transcript.

Because only the text is assessed, pronunciation, intonation and acoustic fluency are not assessed in these results. They are shown as "Not assessed".

Reading, listening and other exercises

Multiple-choice reading and listening questions are scored against an answer key. General learning exercises help you practise. Neither is official exam grading.

Estimates and their limits

Level estimates, such as CEFR levels shown in Progress, are practice estimates based on the work you completed in Parliva, and each one says what evidence it comes from.

Mock exam results are practice results: listening and reading show your answers correct out of the total, and writing and speaking show criterion feedback on the 0–4 scale. They are not converted to the official score scales, CEFR or NCLC levels, because no validated calibration against official results exists for Parliva's mock exams.

AI feedback can make mistakes, and your real exam result may differ. Parliva does not guarantee any score, pass, CEFR or NCLC level.

Start practising free