To turn a lesson dialogue into a video, you write it, voice every line, choose how the pictures look and press Make video in Kukarella. For a one-minute dialogue, the window counts about 38 visuals and about seven minutes of waiting, and you get an MP4 with your voices, a picture for each moment and captions you can switch off. I show every screen below with a real business English dialogue, then how I would use the video in a lesson.
But the video is only as good as your check of it. The pictures are drawn by AI from your lines: a detail can come out wrong, nobody's lips move, and you cannot redraw a single picture in a finished video. Changing the Look or the Shape later makes, and charges for, a new video.
If you want to know first what Make video is, who it is for and how it compares with other tools, read Turn a Dialogue Into a Video: Make Video Explained. This guide is the how.
The whole path in six steps:
- Write the dialogue in DialoguesAI, with lines that say where people are and what they do.
- Voice every line. The video is cut to your voices and keeps your pauses.
- Choose the Look, the Shape and the Captions in Create Visuals. Check the price.
- Press Make video and wait: about seven minutes for one minute of voice.
- Watch it through and check it before anyone else sees it.
- Download it, make a copy without captions, and share the link.
This is the video I take through the steps: The Delayed Pitch, 13 lines, three speakers with three accents, made on 16 November 2026.
Above: Tom is late for a pitch in the rain, Sarah and Aoife move the meeting, and he wins the deal. Photo-real Look, 16:9, captions on.
Step 1: How do you write a dialogue that makes a good video?
Short answer: write it as you would for a listening lesson, but let each line say where the speaker is or what is happening. The pictures are drawn from the words. But a line with nothing to see gives the storyboard nothing to draw, so it repeats or guesses.
Make video starts from a dialogue in DialoguesAI or a voiceover in Text to Voice. Write it yourself, or ask the AI assistant for one from a topic, a level or a file. A video takes up to 10 minutes of voice, but for a class I would stay between 45 seconds and two minutes: short enough to play twice. In 6.9 million viewing sessions on edX, people watched a median of six minutes at most, however long the video was (Guo, Kim & Rubin, 2014).
Give every line something to see. The first line of The Delayed Pitch is "Sarah, can you hear me? I'm at the busy station, and my train's been cancelled. I'm afraid I'm running late." The video shows Tom on his phone in a station hall, under a departures board. The next of his lines, "I'm in the taxi now, but the rain is really heavy," became Tom in the back of a taxi, rain on the window. "You're soaking wet, but the lift is right here" became Aoife pointing Tom to a lift. None of this was in a prompt: it was in the lines.
Tips that help the pictures:
- Name the place early, and again when it changes: the station, the taxi, the office lobby.
- Name the people in the lines: the captions put each speaker's name above their words, and the storyboard keeps each person's look from picture to picture.
- Avoid words the picture would have to show. The storyboard asks for no text, numbers or labels inside a picture. In The Delayed Pitch, Tom's slide shows charts with no words on them. A line like "read the sign" cannot be drawn honestly.
- Keep one moment per line or two. A picture covers a group of neighbouring lines, about one picture for every seven seconds. In The Delayed Pitch, Sarah's "Aoife, Tom's on his way…" and Aoife's "Of course, Sarah…" share one picture of the two women together.
Step 2: Why must every line be voiced first?
Short answer: the video is cut to your voices: each picture lasts as long as its lines do, and your pauses stay. But until every line has a voice and an actor, the window will not offer Make video.
Give each speaker a voice in the editor and press Play to voice the dialogue. Choose the accents your learners need to hear: in The Delayed Pitch, Tom is British, Sarah American and Aoife Irish. Listen once yourself; a wrong stress on a name is easier to fix now than after the video is made.
Then open Create Visuals on the toolbar. Its Video tab shows three steps. The first two are ticks: Written (13 lines, Tom, Sarah and Aoife) and Voiced (all 13 lines, 1:11). If a line has no voice, step 2 says how many are missing and offers Voice N lines from inside the window.
A pause is part of the video. If you want learners to have a beat of silence before an answer, put the pause in the dialogue. The video keeps it.
Step 3: Which Look, Shape and Captions should you choose?
Short answer: choose the Look your learners will take seriously, 16:9 for a classroom screen, and captions on. Nothing is drawn yet, and the window shows the price before you press. But the window starts on the vertical 9:16 shape: switch it if the video is for a screen in the room.
Step 3 says it plainly: "No picture is made yet: choose the Look every picture will be drawn in."
Create Visuals, Video tab: two steps done, and step 3 with the five Looks.Above: the Video tab of Create Visuals for The Delayed Pitch. Steps 1 and 2 are done; step 3 shows the five Looks, each drawn from the same scene.
The Look is the style of every picture:
- Photo-real (the default): real people, real streets. My choice for adults and business English.
- Drawn: flat and friendly, like a textbook. Good for younger learners and for a worksheet that uses stills.
- Storybook: soft watercolour. A story or a reading-aloud lesson.
- 3D cartoon: like an animated film.
- Talking animals: every speaker becomes a different animal. Children love it, but adults may not take a business dialogue seriously this way.
The Shape: Horizontal 16:9 for a projector, a smart board or a shared screen; Vertical 9:16 for learners' phones. The window starts on 9:16. You can only have one Shape per video: the other Shape is a second video, at the same price again.
Captions: on or off. The whole sentence appears near the bottom, the word being spoken lights up in yellow, and the speaker's name sits above it. Make the video with captions on: a copy without them is free afterwards (step 6).
End card: a short card at the end with the video's title and one line about it, with Made with Kukarella. On a paid plan you can switch it off; on a free account it stays, and every frame carries a small Made with Kukarella.com mark.
The Make a video of this dialogue panel: facts line, Look, Shape, Captions, End card and the price.Above: the panel under the Looks. The line under the title is the plan: 1:23, 11 pictures, 1 moving scene, about 7 minutes to make. The price, 42 visuals, sits beside Make video.
Check the price before you press. The small i next to the price opens what it is for. For The Delayed Pitch it said: 11 pictures × 2 = 22, 1 moving scene × 20 = 20, total 42 visuals, and "Voices and captions cost nothing extra." The counts come from the length of your voice, so the price does not change with the Look, the Shape or the captions.
Step 4: What happens when you press Make video?
Short answer: the price is held, and the video is made in five stages while you wait or do something else. But once you press, it runs to the end: there is no stopping it to change a setting.
When you press Make video, the visuals are held, not spent. They are charged when the video is ready; if it fails, they go back to your balance. For a dialogue over three minutes, the window first asks you to check the length, the pictures and the cost.
The window then shows five stages: reading your dialogue and the timing of each line; writing the storyboard; drawing the pictures; making the moving scene; putting it together with your voices and captions. The pictures appear as they are drawn, and the window counts down the minutes left. You can close it: the video waits in Your visuals, and the editor tells you when it is ready.
How long it takes. The window estimates about seven minutes for one minute of voice; the moving scenes are six-second clips made from a picture, about one for each whole minute of voice, none in the first 20 seconds, and they take the longest in a long video. Make the video the day before the lesson, not during the break.
If something fails:
- A picture failed twice: Try again redraws only that picture, at no extra cost.
- A moving scene failed twice: Try again, or Use a still picture, which takes that scene's 20 visuals off the price.
- The picture model refused a picture: the storyboard rewrites it once; if it is refused again, it is left out and the picture next to it stays on screen longer.
- The whole video failed: nothing is spent, and the window says so.
Step 5: What should you check before your class sees it?
Short answer: watch it once from start to finish, as a learner would, with the script beside you. But know what you can fix: captions and the end card for free, a Look or a Shape only by making a new video, and one wrong picture not at all.
My checklist:
- Does each picture show what the line says? In The Delayed Pitch, the taxi line shows a taxi in the rain, and the deal at the end is a handshake in the meeting room. If a picture shows the wrong moment, learners will believe the picture.
- Are the people the same from picture to picture? Sarah keeps her glasses and green jacket from the phone call to the meeting room. If a character changes, learners may think a new person is speaking.
- Ages and genders. A picture can make someone older or younger than the voice. In my French example in the first article, the boy looks younger than he sounds.
- Text that slipped in. The storyboard asks for none, but a shop sign or a screen can still show letters.
- The captions. They show your written lines, timed to the voices. Read the names and any numbers: a typo in the script is a typo on screen.
- The end card. It carries a one-line summary written by the AI. The Delayed Pitch ends with "Despite rain and delays, Tom wins the deal with a great presentation." That gives the ending away: if your learners are predicting the end, stop the video before the last few seconds.
If a picture is wrong, there is no button to redraw one picture in a finished video. You can live with it and say so to the class, which is often a good discussion; change the wording of the line and make a new video, which is charged again; or cut that moment out in your own video editor.
Step 6: How do you download, make a captions-off copy and share it?
Short answer: download the MP4, switch captions off for a free second version, and copy a share link for phones. But the captions-off version replaces the first one, so download the first before you switch.
A finished video in Your visuals, with Download MP4, Share and Change settings.Above: the finished video in Your visuals: the dialogue it was made from, Download MP4, Share, and Change settings.
- Download MP4. A full-HD file named after the video's title, 1920 × 1080 or 1080 × 1920, with your voices as the sound. An ordinary MP4: it goes into a slide deck, a learning platform or a USB stick.
- Make the copy without captions. Go back to the Video tab, switch Captions off and press Make video. The price beside the button reads Free, with "No new pictures are drawn" under it, and it takes a few minutes. It changes the same video, so step 1 comes first. Name the two files so you know which is which.
- Share. Share copies a public link to a page where anyone can watch and download the video without an account. The page shows no price and no project. It is the easiest way to send the video to learners' phones.
Change settings is different: it makes a new video with new choices, and charges for it when you press Make video.
How do you use the video in a lesson?
Short answer: play it at least twice, with a different job each time, and decide the captions for each play. But the research on captions order is mixed, so treat my plan as a starting point, not a rule.
Before playing, show the first picture with the sound off and ask who these people are, where they are and what they might say. Keep it short: British Council trainers advise that a pre-viewing activity "should take no more than ten minutes" (TeachingEnglish).
First play: the gist. One question: what is the problem, and is it solved? Teachers' guides agree on the order: watch once for the overall meaning, then again for the detail (Hughes, 2019; Cambridge, 2026). Do not ask learners to write while they watch.
Second play: the detail. Pause after a few pictures and ask what comes next, or what a speaker means. Then hand out the script.
Captions: on first, or off first?
- What the research found: captions help learners understand at every level, and the extra help fades only at the top, C2 (Pujadas & Muñoz, 2024). Full captions work better than captions of key words only, which learners found distracting (Montero Pérez, Peters & Desmet, 2014). In one study, captions on the first showing helped learners recognise new words by ear, though for learners of Arabic and Chinese the trend went the other way (Winke, Gass & Sydorenko, 2010).
- What many teachers do: captions off first for stronger groups, so they listen, then on to check.
- My plan: A1 to A2, captions on for the first two plays and off for the third; B1 to B2, off, then on; C1 and above, off twice, then on to check the answers.
After the plays: pairs read or act out the dialogue, retell it, or act the scene the dialogue does not show. For a training role-play, play it once, then again with a pause before each answer: "What would you say here?" Then play the line and compare.
Phones: send the share link for homework, or make the 9:16 version for learners who will only ever watch on a phone. In the US, 95% of teenagers aged 13 to 17 have or can use a smartphone (Pew Research Center, 2024); a vertical video fills its screen.
Answer four questions to get a lesson card for your own video: the settings to choose in Make video, the captions for each play, the tasks and the cost. Copy it into your lesson notes.
What does it cost to make one?
Short answer: a one-minute dialogue costs about 38 visuals, and a free account can make one video of up to 45 seconds. But after that one, more videos come with a paid plan.
The price is 2 visuals for each picture and 20 for each moving scene, counted from the length of your voice. The Delayed Pitch, 1:11 of voice, cost 42. Top-ups start at 100 visuals for $12, so that is about $5 of visuals, but the smallest top-up you can buy is the $12 one. The copy without captions is free, and voices and captions cost nothing extra. A second Shape, a new Look or a change of a line is a new video at the full price. The first article has a calculator for any length.
A free account has one free video of up to 45 seconds of voice; a longer dialogue is made from its first 45 seconds, cut after the last whole line that fits. It carries the Made with Kukarella.com mark on every frame.
