The kiss phone AI trend is a roughly 5-second comedy clip in which a person kisses their phone screen and a partner tips a bowl of water over their head. You can make your own version on videosi from one photo of you, with no prompt and no filming.
This guide answers the questions people ask once they have seen the clip: what it is, where it came from, what the AI does and does not draw, and how to get a good result.
Key takeaways
- The kiss phone clip is a kitchen skit: a kiss on the phone screen, then a bowl of water over your head, about 5 seconds long.
- The current wave is emerging: the earliest posts we found are dated 2026-11-17, and no single original creator is confirmed. The idea is older, since Kiss Me Thru The Phone has been a CapCut template since at least September 2026.
- It is a single-person version: only you come from a photo, and videosi never makes two real people kiss.
- A sharp head-and-shoulders photo with even light gives the best likeness. The video has the original sound and no watermark.
- A video costs 10 credits on the Standard tier at 720p. Label the post as AI-generated.
In this guide
- Is kiss phone AI the same as the AI kiss-me filters?
- Why is there a bowl of water in the clip?
- What was the older Kiss Me Thru The Phone template?
- Who is in the video, and which parts come from my photo?
- Why is there only a single-person version?
- Is it safe, and should I label it as AI?
- How do I get a better result?
- What can go wrong in the render?
- How is this different from a CapCut template?
- Which sound does the video use?
- What does it cost and is there a free generation?
Is kiss phone AI the same as the AI kiss-me filters?
No. Kiss-me and kiss-mark filters put a lip print or a puckered overlay on a selfie camera, and that is all they do. The kiss phone clip is a small skit with a story: a person at a kitchen counter, a partner beside them, a phone showing a picture of someone else, a kiss on the screen, a bowl of water, and a soaked walk away. It runs about 5 seconds and the humour comes from the reaction, not from the kiss. videosi makes the skit, so what you get is a short scene with you as the star rather than a filter laid over your face.
Why is there a bowl of water in the clip?
The water is the punchline. Kissing a phone that shows someone else is the setup, and the partner tipping a bowl over your head is the consequence. The beats are easy to follow: you hold the phone in a bright white kitchen, lean in and kiss the screen, the partner lifts the bowl, the water lands, a quick cut shows the phone screen and the shocked reaction, and you wipe your face and back away from the counter. Because the gag is a single clear action, the clip is easy to follow in a few seconds.
What was the older Kiss Me Thru The Phone template?
Kiss Me Thru The Phone has been a CapCut template since at least September 2026, so the idea predates the current wave by a couple of months. Kissing the screen or the camera lens as a selfie joke is older still, from lipstick prints on a lens to kiss-mark filters on TikTok. What changed in November 2026 is the skit format: the earliest posts we found are dated 2026-11-17, say the video is AI generated and point to a CapCut Pioneer template. We list the stage as emerging and no single original creator is confirmed, so we do not credit anyone.
If you want the full beat table and a sample clip, the kiss phone AI page has them.
Who is in the video, and which parts come from my photo?
Only you. Your photo is the single source for your face, hair, skin tone and clothes, and the model keeps them the same in every frame. The partner in the kitchen and the picture on the phone come from the reference clip, not from anyone's photo. That is why the hair and clothes in your picture matter: they are the ones that lean over the phone and get soaked. Choose a photo where you are already dressed the way you want to be seen.
Why is there only a single-person version?
It is a deliberate choice. videosi does not make videos of two real people kissing, so it does not take a second photo and turn it into the person on the phone screen. Keeping the clip to one real person also keeps the joke intact: the kiss goes to a screen, not to another human, and the partner is part of the skit. If you upload a group photo or a picture with a second face, the model cannot tell which person is meant to lean in, so use a photo of yourself alone.
Is it safe, and should I label it as AI?
Yes to both questions, with a note. Only upload a photo of yourself or someone who has agreed to appear. The video is generated, so label it as AI-generated when you post it, as the TikTok templates behind this wave do. The clip is obviously staged, and a clear AI tag avoids any confusion about whether the kitchen scene really happened. The kiss itself is aimed at a phone, which is also why the clip stays light.
How do I get a better result?
Start with the photo, because the skit is filmed from about the chest up. A sharp, front-facing picture of one adult with face and shoulders in view matches best. Use even daylight from the front and avoid beauty filters or heavy colour grading, which drift in the video. Choose the hair and clothes you want to see soaked. Then pick the format: 9:16 for TikTok, Reels and Shorts, or 1:1 or 16:9 for other places. The tier you choose changes the model used, and a clean photo helps more than any setting.
What can go wrong in the render?
Likeness is not guaranteed, and the water splash covers your face for a moment, which can soften it for a few frames. Kissing a screen is a tiny motion, so the lean-in can come out slightly early or late against the splash. The kitchen, the water gag and the timing follow the original clip and cannot be changed. The water and the reaction are drawn by the AI model, so they look like the original but are not frame-identical. A failed, timed-out or blocked render returns its credits automatically.
How is this different from a CapCut template?
A CapCut template swaps clips and effects inside an editor, and the people in the finished video are whoever you drop in. videosi works from a single photo: the model reads your picture, follows the motion of the reference clip and renders a new 5-second video with you as the person who kisses the phone. You do not film anything, and there is no prompt to write. The trade-off is that you follow the original skit rather than editing your own cut.
Which sound does the video use?
There is no single confirmed trending track. The clips in this wave use the skit's own original sound, and the videosi video comes with the reference clip's sound already on it, so you do not need to find or add audio. If you want a different song, replace the sound in the editor you post from. Because the clip is only about 5 seconds, the original sound is short and usually works best left as it is.
What does it cost and is there a free generation?
A video costs 10 credits on the Standard tier at 720p. The Single pack is 10 credits for $8.90, so one video is one pack. There is no free generation, and the higher quality tiers cost more credits per video. The roughly 5-second MP4 has no watermark. If a render fails, times out or is blocked, the credits come back automatically, so you pay for finished videos only.
Ready to try it? Open the kiss phone AI generator, upload one clear photo of yourself and press Generate. For other phone trends, see the iPhone interaction effect, the metro phone trend and what is trending this week.



