PREDICTIVE ENGLISH ASSESSMENT

Know your English level
in 50 minutes

One test. Four estimates. Take a calibrated grammar, vocabulary, reading and listening test plus AI-graded writing and speaking sections, and get your predicted CEFR level, IELTS band, TOEFL iBT and TOEIC scores instantly.

Start the free test How scoring works

Free score prediction · Optional verified certificate for $9.00 (about Rp 145.000)

How it works

1

Take the test

34 calibrated multiple-choice questions across grammar, vocabulary and reading, a device-spoken listening section, 2 writing tasks and 4 spoken tasks.

2

AI grades writing & speaking

Your essays and spoken answers are evaluated by the DeepSeek AI examiner against IELTS-style rubrics: fluency, coherence, lexical resource and grammar.

3

Get your prediction

Weighted difficulty scoring with a guessing correction converts your performance into CEFR, IELTS, TOEFL iBT and TOEIC estimates with per-skill breakdown.

4

Share a certificate

Optionally buy a verified certificate for $9.00 via Duitku or PayPal. It comes as a shareable PNG on a unique verification URL, ready for LinkedIn.

Four scales, one result

Your single performance is mapped to the scale employers and universities ask for.

CEFR

A1 - C2

Common European Framework level, including plus levels (B1+, B2+, C1+).

IELTS

1.0 - 9.0

Overall band plus listening, reading, writing and speaking sub-bands.

TOEFL iBT

0 - 120

Section scores for reading, listening, writing and speaking.

TOEIC

10 - 990

Listening & Reading scores, the format most employers request.

Methodology

Calibrated items. Every question targets a CEFR level (A2 to C2) and carries a difficulty weight from 1 to 5. Higher-level items count more, so guessing a few easy questions cannot place you at C1.

Guessing correction. Four-option multiple choice yields about 25% correct by blind guessing. We subtract a 18% baseline and rescale before mapping to levels.

AI writing & speaking grading. Writing and speaking are graded 0-9 on IELTS-style rubrics by the DeepSeek AI examiner. Speaking responses are transcribed in your browser and assessed on fluency, vocabulary, grammar and content. If the AI service is unavailable, a conservative offline estimator (capped at band 6.0) is used and the result is clearly flagged.

Scale mapping. Combined score S = 0.60 × adjusted MCQ accuracy + 0.40 × writing band / 9 (or 0.50 + 0.25 + 0.25 when speaking is submitted), then S is mapped: below 27 A2, 45-53 B1, 63-71 B2+, 81-89 C1+, 90+ C2. IELTS, TOEFL and TOEIC values are derived from per-skill adjusted accuracies.

Honest disclaimer: LinguaProof gives a predictive estimate based on a single sitting. It is not an official IELTS, TOEFL or TOEIC score and carries no guarantee of the official result, though the item design follows the same skill and level structure.

Pricing

Score prediction

Free

  • Full 50-minute assessment
  • CEFR level + IELTS / TOEFL iBT / TOEIC estimates
  • AI writing & speaking feedback
  • Answer review with explanations
  • Unlimited retakes

Verified certificate (optional)

$9.00 · Rp 145.000

  • Digital certificate in PNG format
  • Unique public verification URL
  • One-click LinkedIn / X / WhatsApp share
  • Anyone can verify the ID at /verify
  • Pay with Duitku (QRIS, VA, e-wallet)
Get started

Payment gateways run in demo mode until merchant credentials are configured.

FAQ

Is this an official test?

No. LinguaProof is a prediction tool. Its items mirror the skills and CEFR levels tested by IELTS, TOEFL iBT and TOEIC, and the score mapping follows published CEFR alignment studies, but it is not affiliated with any official testing body.

How long does the test take?

Plan for about 50 minutes: roughly 20 minutes for the multiple-choice sections, 10 minutes for listening, 15 minutes for the two writing tasks and 5 minutes for the speaking section.

Do I need a microphone for the speaking section?

Ideally yes - speak your answers and the browser transcribes them for the AI examiner. If your browser cannot do in-browser speech recognition (for example Firefox), the speaking section switches to a typed fallback automatically.

How accurate is the prediction?

With honest, unaided answers the CEFR placement is typically within one sub-level (e.g. B1+ vs B2). Writing grading depends on the AI examiner; when it is unavailable the offline estimator is intentionally conservative.

What do I get for $9.00?

A digital certificate as a shareable PNG on a unique verification URL, plus one-click LinkedIn sharing and a public verifier so anyone can confirm the ID.