Spoken checks for university courses
Hear how every student thinks.
Aloud gives each student a short spoken viva on what you taught. They explain their reasoning out loud, an AI examiner asks the follow-up, and you see, topic by topic, who understood.
For programming, maths, science and humanities courses. Built for classes of 20 or 500.A written answer no longer tells you who understood it.
Code, essays and problem sets can be produced without being understood. A one-on-one viva still shows the difference in a few minutes, but no one has time to run one for every student in a 400-person course.
Aloud runs that conversation for you, for every student, on the material you choose. You stay in charge of the questions and of every result.
How it works
From lecture notes to a class picture, in four steps
- 01Upload what you taughtLecture notes, slides, readings or code. Aloud finds the ideas students should be able to explain.About 2 minutes
- 02Approve a pool of questionsQuestions apply the idea to something new. A second AI pass checks every answer key; you approve.About 10 minutes
- 03Every student talks it throughEach student gets their own draw from the pool and explains out loud. The examiner asks a follow-up when it matters.About 15 minutes each
- 04See who understoodResults per topic and per student, with the words behind each one, as soon as students finish.Live
What you get
Not a grade. A picture of what your class understands.
- Every topic, every student: understood, partly, or a gap.
- The exact words behind each result, so you can check it.
- Which questions the class found hardest, and the mix-ups behind them.
- Signals worth a look, like long silences. Never automatic verdicts.
- You can change any result with one click.
Any subject
Questions about your material, in new situations
Students don't recite your slides. They apply the idea: a new program, a new case, a new objection.
Computer science
This never finishes for n = 5. Find the bug and explain why.
def fact(n):
return n * fact(n - 1)
Economics
A town has one big employer. Could a higher minimum wage raise employment there? Walk through why.
Listen for: monopsony, wage below marginal product, labour supply
Philosophy
A city makes motorcyclists wear helmets. Would Mill accept this law?
"His own good, either physical or moral, is not a sufficient warrant."
Biology
A red blood cell is dropped into pure water. What happens to it, and why?
Listen for: osmosis, water moves into the cell, no cell wall, it bursts
Built for real courses
Large classesSections, hundreds of students, search and filters. Built for a 400-student intro course.
Different questions per studentEach student draws from a pool, so answers can't be passed around.
Weak wifiRecording keeps going if the connection drops. Students pick up with a fresh question.
AccommodationsExtra time, questions read aloud, or typed answers, confirmed by you.
You approve every questionNothing reaches students until you do, and you can rewrite anything.
Your class listImport it once. Students sign in with their university email.
Student data
Recordings belong to your course, not to us
- Only you and your teaching assistants see transcripts and recordings.
- You choose how long recordings are kept. They’re deleted automatically after that.
- Student answers are never used to train AI models.
- Students agree to recording before they start, and can ask for accommodations.
“One hundred percent of professors are experiencing an existential crisis regarding AI assessments and would be highly interested in helpful technological solutions.”
“Such a tool would solve the challenge of providing individualized attention in large classes by zeroing in on specific student and class weaknesses.”
“Teaching has become significantly more difficult and highlighted the ongoing challenge of determining the exact extent of AI use in student submissions.”
Run your next lab as a viva.
Set up a course in a few minutes. We'll help you with the first assignment.