Set for junior high: A2–B1 — textbook units and exam content turned into real speaking, with a runway for the silent ones.
Hungry deals smart conversation cards: level-matched questions with a scaffolded runway to answer, from your topic or your teaching material. It plays in any browser — one device per student or pair (school tablets, Chromebooks, iPads, or a teacher-approved phone); a projector or printed cards work as backups. Nothing to install.
Six real classroom situations. Each one: what to do, and a live button that opens the app already set up for it.
The eight habits that make it work — from the teachers already using it.
60 free cards with a Google sign-in, then 3 a day. Have a code from Spedur or Unison? It unlocks everything.
A speaking activity where the same five hands always go up. Hungry deals every student a question at their level (Pre-A1–C2 — pick one level or two adjacent as a range), with a reply frame and example answers, so there's always a runway to speak.
🔬 What research has found: Without a structure, many students in group work say very little. So the cards deal the turns, and every student answers every card.
Upload a worksheet, PDF, Word doc — or photograph textbook pages straight from the app camera, several at once. The cards are generated from that material: teach what you teach, practice what you teach.
🔬 What research has found: Recalling words to use them is one of the most studied ways to make learning stick. So the questions come from what you're teaching this week.
Some teachers get to roam; some are measured by the scores their students bring back, on content that is already decided. When the 段考 range is set, wandering questions waste minutes you don't have. Upload the exact material and set strictness to Stick to it — test/review: every question comes only from what you gave it — exam week, or all term if that is your reality. Explore around it stays the default when there is room to roam.
🔬 What research has found: Recalling what's on the test, instead of reading it again, is one of the most studied ways to make it stick. So strictness keeps every question inside your material.
Every scaffolded card ends with + my own answer. A student types what they actually want to say — in any language — and it comes back fitted to the reply frame at their level, already selected — and when what they mean can't fit the frame, the frame itself is minimally revised, with the changed words shown in red. It's saved: it will be there on that question next time, and it sharpens what the cards ask them about.
🔬 What research has found: Using a new word in a sentence of your own helps it stick. So a student's own answer comes back as a sentence they can say.
Social mode deals warmer, more playful cards — for school clubs, English cafés, church groups, family tables. Same engine, different dinner.
🔬 What research has found: People remember language better when they connect it to their own lives. So the cards ask about the players themselves.
▶ Try Social mode ↗ Opens in Social · Bilingual — both languages on every card.Set the level and topic, print the cards, and run the whole activity on paper. Any setup can also travel as a QR code — students scan it once and practice the same deck on their own device at home.
The exact printing taps are just below in Try these features ↓
Most speaking apps put one student in front of a screen. Hungry puts two to five students in front of each other, and the screen only holds the help.
This matters for three plain reasons. First, a child says more when a real person is waiting for the answer. Second, when two students work on the same question they can go further than either one alone — one knows the word, the other has the idea. Third, new ideas come out of people talking, not out of one person answering a machine.
The screen gives the question, the frame and the words. The talking belongs to the children.
You do not learn to speak by listening alone. You learn by speaking.
When you say something out loud, your brain must find the words, put them in order, and get them out — now, to a real person who is waiting. Reading and listening never force you to do that. Merrill Swain gave this idea its name in 1985: taking language in is not enough, you have to push it back out.
The second part is what happens when it goes wrong. You say something. Your partner does not understand. They ask again. You try other words. Michael Long argued in 1996 that much learning happens right there, in the repair. Being honest: some studies support this and some do not. It is a real effect, not a sure thing, and only one part of learning a language.
So why a classmate, and not the teacher? Jenefer Philp, Rebecca Adams and Noriko Iwashita (2014) say it plainly: teachers are teachers and peers are peers. They offer different chances. When the teacher asks the class, one child answers and the rest wait. With a classmate, you talk the whole time. And students are more willing to say “I don’t understand” to a classmate than to a teacher. “I don’t understand” is the door. A child who will not say it is a child the lesson passes over.
What Hungry does about it: two to five students sit facing each other, turns are dealt in order so every child speaks, the cards come from the teacher’s own material, and the help sits on the screen where they can hide it when they are ready.
Three plain things happen when a child explains something to another child.
First, they must pull it out of their own head. Henry Roediger and Jeffrey Karpicke showed in 2006 that students made to recall something remembered far more two days and a week later than students who read it again — even though reading again felt easier.
Second, they must say it to someone. Logan Fiorella and Richard Mayer found in 2013 that students who actually explained a lesson still had a clear advantage a week later. Students who only prepared had almost none. Preparing is not the same as speaking.
Third, it comes back. Nicholas Cepeda and colleagues (2006) reviewed hundreds of experiments: practice spread over time beats the same practice crammed together.
Does talking with classmates work? The best answer is a 2020 review by Harriet Tenenbaum and colleagues, pooling 71 studies of children aged 4 to 18. Peer interaction did better than the other ways of learning it was compared with — best when the children had to agree on an answer, not just take turns.
Now the honest part, from the same study: peer interaction was not better than one adult sitting with one child. So this does not replace a good teacher. It is how you give every child time to talk when one teacher cannot sit with each of them.
What Hungry does about it: the card asks a real question the child must answer out of their own head, everyone speaks, the words come back over the following weeks, and it sits on top of the teacher’s lesson — never in place of it.
Kevin Dunbar spent a year inside working science laboratories, watching how discoveries were made. He reported in 1997 that the weekly lab meeting was one of the main places new ideas were born. One person shows a result. Others ask questions the presenter had not thought of. Alone, Dunbar found, people struggle to think of a second explanation for what they see — in a group, somebody else supplies it. And when a result was strange, the group talked about it far more than when it was expected.
Brian Uzzi and colleagues looked at 17.9 million scientific papers in 2013. Papers written by three or more people put an unusual combination of ideas into their work more often than papers by one person.
Here is a picture to explain it: one person with a torch can only look where the torch points. Two people with two torches look in two directions — and each sees what the other’s light landed on. That is what a conversation does with ideas.
Being straight about it: this research is about scientists, not children in an English class. Nobody has shown that talking in pairs turns children into inventors. It does show that the question you did not expect is what moves thinking — and only another person can hand you one.
What Hungry does about it: the cards come from the teacher’s own material, so the unexpected question lands on something the class actually knows — and a real person sits across the table to ask it.
Hungry is extra speaking practice on top of your normal lessons. It does not replace them; it gives every student far more turns to actually talk.
The exact taps for all of these live above in Try these features ↑
Works in any browser, nothing to install. 60 free cards with a Google sign-in, then 3 a day — enough to feel a real lesson. Your own topics, uploads and photos are included in those 60 cards; event codes and Multiple sessions need a code.
Have a code from Spedur or Unison? Tap ⋯ → Full Access → enter it under “Have a license code?” → Unlock. One team or school code covers all its seats.
Open Hungry ↗