What's new in ToetsTester

A new assessment now starts with a conversation. You describe in plain language what you have in mind - how many questions, which types, which level, where the emphasis should be - and add your own material with the paperclip, such as a chapter, a handout or an older assessment. ToetsTester reads those files and keeps track of the plan while you talk, shown as short labels underneath the chat. Title, subject and level stay ordinary fields, so you can always adjust things yourself. Happy with it? Click Continue and the assessment is generated. Without any material an assessment can be built from the conversation alone. Multiple-choice questions also come out cleaner now: every option stands on its own, with exactly one correct answer.

Preparing your correction model happens in that same chat. Uploading the model and adding the exam questions separately are now a single step: you upload or paste your correction model, and the assessment with the questions can join the same conversation. ToetsTester tells you what it found - how many questions, how many points - and points out on its own when questions have no worked answer yet, because those cannot be graded. Question texts, worked answers, points, question types and numbering you then adjust simply by asking; every change is saved right away. Above the chat you see at a glance whether the model is ready for grading, and View model takes you straight to the editor. Want to start over? One button does that.

Every assessment now has a "Digital assessment" option. You set the window in which the test is open, how much time each student gets and which instructions appear at the top. You pick participants as a whole class at once or as individual students, and you decide whether open answers should be graded by AI - closed questions are always graded automatically. Students with an email address receive an invitation with their own link and verification code; for the others you copy their personal code straight from the screen.

Students open the link, enter their access code and take the test in a calm, distraction-free screen: question by question, with the points shown. Answers are saved automatically as they work and stay on the device if the connection drops, so nothing gets lost. Before submitting, students see how many questions they left unanswered.

While the assessment runs, the overview shows per student whether they still have to start, are working or already submitted, which question they are on and whether anything went wrong. You can grant someone ten extra minutes or submit an attempt on their behalf. As soon as a test comes in it lands in your usual grading screen - paper and digital assessments now sit side by side in one overview per assessment.

Every assessment can now have a practice zone. You choose which material counts as a source - the assessment with its answers, or your own uploaded files - and which question types you want: open questions, multiple choice, true or false, multiple correct answers or applied questions. ToetsTester recognises the topics in your material (you can add your own too) and builds a set of sample questions from it. You review those questions with their model answer and explanation, edit them where needed, and mark per question whether it must be included or left out. Extra AI instructions let you steer tone, level and focus.

Students enter the practice zone through their own access link. They pick a topic, work through a round of questions and get immediate feedback with the correct answer each time. When things go well the next questions gradually get harder, and any question they got wrong can be reviewed later. They can generate new questions as often as they like, and a streak of correct answers keeps practising rewarding.

The Students tab shows you per student how much they practised, how accurate their answers are and how they are progressing per topic. ToetsTester automatically writes a summary with strengths and points for improvement, so you quickly see where someone still needs help. Did a student not hand in an assessment? You can still give them access to the practice zone with just an email address.

Until now ToetsTester could only assess text, which meant questions asking students to draw something stayed manual work. In the correction model editor you can now pick the "Image answer" question type and upload a reference image of the expected answer. Your assessment criteria and maximum score still count, and the image is added as extra guidance.

While grading, ToetsTester locates the drawn answer on the scanned page and compares it with your reference image. As long as that is running you can see which question is still being assessed; the grade only appears once everything is done. Sometimes it does not work out, for example when the answer could not be found on the page. That single question is then clearly marked as "manual review needed", along with the reason. The rest of the assessment is graded as usual and you only fill in those points yourself.

Your work revolves around the assessment, and that is now reflected in the app. Your dashboard has a "Recent assessments" section you can search through and collapse or expand with one click — just like your recent classes. From there you open an assessment in its own overview, showing the student portals that belong to it and how grading is progressing. You can edit an assessment title right on that page, separately from the correction model title.

When grading scanned assessments you no longer have to hunt through the PDF yourself. As you scroll through the questions, the document follows along to the matching student answer. You can also click a question number to jump straight to that answer in the PDF.

You switch this automatic navigation on or off in the settings menu with the "Automatically scroll" option. It becomes available once the answer positions in the PDF are known.

Until now ToetsTester calculated grades with a fixed N-term formula. From now on you decide which norm best fits your assessment.

You can choose from six methods: points per grade point (counting correct answers), points per mistake, steering towards a target class average, the classic N-term, an N-term with a cutoff, and a score table where you link a grade to each range of points yourself. As you adjust the settings you immediately see the effect on the grades and the class average.