Menu
Start a project

Product · Under construction

AI language and examination platform

Under construction. This page describes what is being built rather than something that can be used today.

Status: Under construction.

Abstract drawing of a sheet of question lines beside a dashed circle holding a check mark, standing for a generated exam and its evaluation.

What is being built

A platform for language testing in which a language model generates the examination itself: the questions, the expected answers, and the evaluation of a candidate's response, across difficulty levels and in a structure similar to established language certification exams.

The first phase is text-based. A candidate reads and writes; the model produces the material and assesses the response against the level being tested.

Phase two: audio

The second phase adds audio testing: listening comprehension, spoken responses, and evaluation of speech. That work has not started, and the page will say so until it has.

What is not decided

The model, the provider and the AI architecture are still being chosen. That choice depends on how well each option performs on the evaluation material once it exists, and on where candidate audio is allowed to be processed. Publishing a stack before those questions are answered would be guesswork.

Architecture

The evaluation set comes first

A generated exam is only useful if its difficulty is trustworthy. The measurement material is being built before the feature, not after it.

Structured output

Questions, expected answers and evaluations are stored as structured records rather than as generated prose, so they can be compared, reviewed and corrected.

Technologies

Python Node.js LLM APIs open-weight models

Tell us what you are building.

Write in English, French, Romanian or Arabic. Every message gets a written answer within one business day.