Which Tool Gets Your Class Debating Out Loud
Kialo, Dlibr and NOVA do not do the same thing. The dividing line is not price: it is written or spoken, deferred or live. An honest comparison.
By ELBO
The question almost always arrives in the same order. You want your class debating, you look for a tool, and you find three or four names that appear to do the same thing. They do not.
The dividing line is not quality. It is written or spoken, and deferred or live. Settle those two questions and the choice mostly makes itself — and it is very common to need two tools rather than one.
The two questions to settle first
Do you want writing or speaking? A written debate leaves a readable, re-readable, annotatable trace. Everyone takes the time to phrase things properly. It is excellent for building reasoning, spotting a fallacy, learning to structure. A spoken debate produces something else: the immediate reaction, the hesitation, the rephrasing under pressure. These are different skills, and they are not trained with the same tool.
Live, or spread over time? An asynchronous exchange lets everyone take part without waiting their turn — the shy student writes at 10pm, the slow thinker mulls it over for two days. A live exchange imposes the real-time constraint, which is precisely what you want to train when an exam, an interview or an oral defense is coming.
The comparison
| Kialo Edu | Dlibr | NOVA | |
|---|---|---|---|
| Format | Written | Written | Spoken (voice) |
| Pace | Asynchronous | Not stated on their page | Live |
| Structure | Argument tree, pros and cons under each point | Arguments linked by "because", "on the other hand", "therefore" | Two sides, timing, speaking turns |
| Who you argue with | Other students | Other students | An AI that pushes back, or another student |
| Student accounts | Not stated on their home page | Yes — separate admin, teacher and student interfaces | No: a code and a first name |
| Price | Free, stated as permanent | Not disclosed | Free to start |
| Origin | International, English-first | Quebec, French and English | Quebec, French-first |
Sources: the official Kialo Edu and Dlibr pages, consulted on 23 September 2026. What was not stated there is marked as such.
When each one is the right answer
Kialo Edu is hard to beat for making the structure of a disagreement visible. The tree literally shows where an argument branches and which objection attaches to what. For a course whose goal is to map a controversy — philosophy, history, citizenship — it is the tool that makes reasoning visible. And it is free, with no asterisk.
Dlibr comes out of the Quebec college network and imposes three connectors — "because", "on the other hand", "therefore". That formal constraint is the point: it forces students to name the logical relationship between two ideas, which free writing lets them dodge. Its visualisation and assessment tools are built for the teacher who has to grade, not only to facilitate.
NOVA does neither. It gets students talking, live, against an AI that answers and pushes back. Its reason for existing fits in one sentence: in a class of thirty, a conventional oral debate gives each student two minutes — a full hour, and in practice twelve students who speak and eighteen who watch. In parallel rooms, everyone debates at the same time.
What changes when the debate is spoken
Three settings exist because a French-as-a-foreign-language teacher in Hanoi asked for them, for B1–B2 students who found the AI too fast: speech rate (four levels, 0.7 to 1.15), answer complexity (three levels, calibrated against the CEFR) and on-screen text while the AI speaks — reading while listening roughly doubles comprehension in a second language.
At the end, if the teacher enabled it, a session report returns the transcript, a summary, and an indicative pre-assessment out of twenty — arguments 6, listening 4, rebuttal 4, language 4, format 2. It is labelled "for the teacher to confirm", and that label is not decorative: a machine does not know that a student has improved enormously since last month.
What to check before choosing, whichever tool it is
- Do your students have to create an account? That is the first friction, and often the only one that matters. In a class of minors it is also a data protection question.
- Where does student work go? A recorded debate is personal data. Ask how long it is kept and who can see it — the answer should be written down somewhere.
- Is free actually free? Some tools are free without conditions, others are free up to a threshold.
- What survives if the tool shuts down? Can you export the work?
The honest answer
If you are teaching written argumentation, Kialo or Dlibr will do it better than NOVA, and there is no reason to deny yourself. If your problem is that your students do not speak — and that is the most common problem in a language classroom — writing will not solve it, however good the tool. Mouths have to open.
Plenty of teachers use both: writing to prepare and structure, speaking to defend. That is probably the right answer.
Going further: if the problem is a class that stays quiet, what actually gets a silent class talking. If it is fear of the debate collapsing, how to structure one so it holds. If it is marking, grading class participation without losing your evenings.
Free, no sign-up: a ready-to-teach 60-minute debate lesson, a bank of 100+ topics sorted by level, and an oral defense procedure for the age of ChatGPT.
Read next
How to Get a Silent Class to Talk
Classes rarely go quiet out of laziness. They go quiet because being wrong is expensive in front of peers, because three students hold most of the airtime, and because teachers wait one second where five are needed. What actually changes it, and what does not.
EducationGrading Class Participation Without Losing Your Evenings
Participation is the least reliable grade a school awards: it runs on memory, and memory favors the students who take up room. Here is how to make it defensible — six to eight students per session, a four-criterion rubric filled in during class, and a worked 20-point example.
EducationHow to Run a Classroom Debate That Doesn't Go Off the Rails
Fear of a debate collapsing stops many teachers from ever running one. But debates rarely fail because of the topic — they fail for lack of structure. Random sides, mandatory restatement, stepping out of role, topics to leave out, and what to do when it goes wrong anyway.