Overview
A test lives inside a course section. Students take it as part of their learning flow — the timer ticks, answers are auto-saved, and results appear immediately on submission. Objective questions are graded the moment the student submits. Subjective questions (short text, long text, file upload) are sent to the AI grader in the background — results typically appear within 30 seconds.
To add a test: open the curriculum builder, expand any section, and click + Add test.
Test settings
When you create or edit a test, the test editor shows these settings:
| Setting | Description |
|---|---|
| Title | Shown in the curriculum sidebar (e.g., "Unit 1 Quiz" or "Mock Exam 3"). |
| Kind | QUIZ (ungraded / practice check), EXAM (formal, tracks attempts), or PRACTICE (unlimited attempts, no pressure). |
| Duration (minutes) | Set a time limit. The student sees a countdown; the test auto-submits at 0. Leave blank for no time limit. |
| Max attempts | How many times a student can attempt this test. Leave blank for unlimited. Attempts where the student ran out of time still count. |
| Pass score (%) | Optional. If set, the results view shows a green "Passed" or red "Failed" banner. |
Question types
After saving the test settings, click + Add question to open the question type picker. Seven types are available:
| Type | Graded by | Best for |
|---|---|---|
| MCQ (single correct) | Auto (instant) | Recall, definitions, formulas |
| Multi-select | Auto (instant) | Questions with multiple right answers |
| True / False | Auto (instant) | Quick fact-checks, misconceptions |
| Fill in the blank | Auto (instant, per-blank) | Vocabulary, equations, key terms |
| Short text | AI (async, ~30 s) | 2–3 sentence answers, definitions |
| Long text | AI (async, ~30 s) | Essay-style, explanations, analysis |
| File upload | AI via OCR (async) | Handwritten work, diagrams, scanned solutions |
MCQ — single correct answer
The student picks exactly one option from a list. You define which option is correct.
To create:
- Enter the question text.
- Click + Add option to add answer choices. Add at least two; four is typical for exam-style questions.
- Click the radio button next to the correct option to mark it as correct.
- Set the point value (default: 1).
- Optionally add an Explanation — shown to the student after submission.
Multi-select
The student selects all options they believe are correct (checkboxes, not radio buttons). The full answer is correct only when the selected set exactly matches all correct options.
Grading: Full points for a perfect match; zero for any mismatch. Partial scoring is not applied — the full set must match.
To create: Same as MCQ, but toggle the checkboxes next to all correct options. Multiple correct options can be marked.
True / False
A statement that is either true or false. The student sees two large buttons — T and F.
To create: Write the statement in the question text. Toggle the correct value radio (True or False). Set points.
Fill in the blank
The question text contains one or more blanks. Each blank is answered with a short phrase. Every blank is graded independently; the question's total score is the sum of all blanks.
To create:
- Write the question text, putting
[blank]as a placeholder wherever you want a blank field (the UI handles this as a visual hint — the actual blank fields are defined below the text). - In the Blanks section, each blank has:
- Accepted answers — comma-separated list of valid answers (e.g.,
Newton, Sir Isaac Newton, Isaac Newton). - Case-sensitive toggle — off by default. Turn on if capitalisation matters.
- Points — per-blank point value.
- Accepted answers — comma-separated list of valid answers (e.g.,
- Click + Add blank to add more blanks to the same question.
Short text
A text field expecting a 1–3 sentence response. AI grades it against your rubric.
To create:
- Enter the question text.
- In the AI Rubric / Model answer field, describe what a correct answer looks like — key points, expected facts, or paste a model answer. The more specific you are, the more accurate the AI grading.
- Set the max points for the question.
Students type their answer in a multi-line text box. The AI evaluates the content, awards a score proportional to how well key points are addressed, and generates brief written feedback.
Long text
Identical to short text, but the student gets a larger text area and the expectation is a fuller, essay-style response (a paragraph or more).
The AI grades against your rubric, awards a score, and provides paragraph-level feedback (strengths, gaps, suggestions). Long text questions are well-suited for 5–15 mark analysis or explanation questions.
File upload
The student uploads an image (handwritten answer, diagram, scanned work). The platform runs OCR to extract text from the image, then passes the extracted text to the AI grader along with your rubric.
To create:
- Enter the question text (what you want them to write / solve).
- Optionally enter a Reference answer URL — a URL to a PDF or image of a model solution. The AI can use this as a reference when evaluating the student's submission.
- Set a rubric and max points.
On submission, the student pastes or types a URL to their uploaded image (students upload the image to any public URL — Google Drive, Imgur, or similar, then paste the link). The OCR step runs automatically.
Rubrics for AI grading
A rubric is the instruction you give the AI about how to evaluate an answer. Good rubrics produce accurate, consistent, and fair scores.
Examples of effective rubrics:
- For a 5-mark short text: "Award 1 mark each for: (1) correct definition of osmosis, (2) mention of semi-permeable membrane, (3) direction of water movement (high to low concentration), (4) real-world example, (5) correct use of the term solute/solvent."
- For a 10-mark long text: "Award up to 4 marks for correctly identifying the three Newton's Laws; up to 3 marks for applied examples; up to 3 marks for clear explanation and logical flow. Deduct 1 mark for each factual error."
Auto-grading vs AI grading
The platform uses two distinct grading paths:
- Auto-grading (instant): MCQ, multi-select, true/false, and fill-in-the-blank are graded the moment the student submits. No AI involved — pure deterministic logic. The score appears in the results view within milliseconds.
- AI grading (async, ~30 s): Short text, long text, and file upload are sent to the AI grader after submission. The student sees their objective score immediately and a "AI grading in progress…" banner for subjective questions. The full score updates within ~30 seconds.
Creators can view and override AI-awarded scores from the assignment/test grading panel at any time. See the AI Grading & OCR guide for how to review and override AI grades.
Editing & deleting questions
In the test editor, each question in the list has a pencil (edit) and trash (delete) icon.
- Editing reopens the question editor. All fields are editable — including the correct answer. Changes apply immediately.
- Deleting removes the question permanently, along with all student answers for that question across all attempts.
The question list also shows the total points the test is worth — the sum of all question point values — which updates live as you add or edit questions.
Tips for effective tests
- Mix question types. A test with 5 MCQs, 2 fill-in-the-blanks, and 1 short text is more diagnostic than 8 MCQs.
- Always add explanations to MCQ/TF. Students learn from mistakes. The explanation field is shown after submission and is one of the highest-engagement moments in the learning flow.
- Keep time limits realistic. Budget 1–1.5 minutes per MCQ and 5–8 minutes per short-text question. Rushing students produces noise, not signal.
- Set a pass score for structured courses. A pass threshold (e.g., 70%) gives students a clear target and makes completion meaningful.
- Use Practice kind for warm-ups. A practice test with unlimited attempts at the start of a topic lets students self-assess without the pressure of attempt limits.