The ACP Pradyuman AI trend is a set of short AI comedy clips in which the stern officer from the Indian TV series CID sits straight-faced and then cracks into a laughing fit, often captioned 'ACP aura -999999'. You can make the sofa version from one photo of yourself on videosi, as a 10 second video with the original sound on it.
This guide goes deeper than the trend page FAQ: who the character is, why the joke works, what the aura caption means, why faces drift in a laugh, which photo to use and how to post it responsibly.
Key takeaways
- The ACP Pradyuman AI trend is a format, a stern face that collapses into laughter, not a single song or creator.
- The joke works because of contrast: four calm seconds, then a laughing fit, then a repeat laugh outdoors.
- One sharp, front-facing photo of one adult with a closed mouth gives the best likeness.
- Faces can drift during laughter, so a better photo helps more than extra tries.
- Use only your own photo or a friend's with consent, and label the video as AI-generated.
In this guide
- Who is ACP Pradyuman and why is he the face of this trend?
- Why does a straight face turning into a laughing fit work as a joke?
- What does "ACP aura -999999" mean?
- Where did the trend start and how big is it?
- Why does the face drift during a laughing fit?
- Which photo gives the best result?
- How does a single photo become the full 10 second clip?
- Why do many of the original clips carry a Kling AI mark?
- How is this different from a filter or a CapCut template?
- Should you label the video as AI, and how?
- Is it okay to make this with a real person's likeness?
- What does one video cost?
Who is ACP Pradyuman and why is he the face of this trend?
ACP Pradyuman is the stern senior police officer from the long-running Indian TV crime series CID. He is a fictional character, and ACP stands for Assistant Commissioner of Police, his rank on the show. On the show he is the calm, serious boss who rarely smiles, speaks in clipped lines and treats every case with total gravity. That reputation is the whole setup for the trend.
The AI clips take that image and drop it into ordinary life: a night market, an elevator, a sofa with a plate of biryani, a table with an energy drink. Because the audience already knows the character as the man who never loses his composure, any slip into laughter, panic or embarrassment reads as a joke before a single caption appears.
Only the character's name and the joke format are used on the trend page. The page does not claim that anyone real said or did anything, and your own photo replaces the man in the clip.
Why does a straight face turning into a laughing fit work as a joke?
Comedy often comes from contrast, and this clip has a clean one. For the first four seconds the man sits upright in a black suit, stiff and unamused, while a hand offers him a plate. The viewer is waiting for him to react the way the character always reacts: with a stare.
Then he cracks. He bends forward, slaps the cushion with both hands and shakes with laughter. The bigger the gap between the serious setup and the helpless payoff, the funnier the cut. That is why the clip does not rush the first beat: the stiff pause is what makes the laugh land.
What does "ACP aura -999999" mean?
"Aura" is internet slang for a person's cool, status or presence. People give themselves or others aura points for impressive moments and take them away for embarrassing ones. A number like -999999 is deliberately absurd: it says the character has lost all of his composure, far past any normal penalty.
On these clips the caption sits over the moment the stern officer breaks into laughter. The idea is that a man known for unshakable authority has just burned through every point he had. The caption mixes English internet slang with the Hindi and Urdu humour of the creators who post it, and the hashtags are #cid and #acppradyuman.
videosi does not render captions into the video, so if you want the joke in your own clip, type 'aura -999999' in TikTok or CapCut over the laughing fit around the four second mark.
Where did the trend start and how big is it?
The earliest clips we found are dated 24 October 2026. By 13 November 2026 one clip captioned 'ACP switched the mode' had passed 9 million views on TikTok, with several others between 1 and 3.5 million. The trend is posted mostly by Indian and Pakistani creators.
We have not confirmed a single original creator or a first post, so we do not name one. We also found no single trend sound: each clip uses its own original audio, usually laughter and a short spoken Hindi line, which TikTok lists as 'original sound'.
Why does the face drift during a laughing fit?
A laugh is one of the hardest expressions for a video model to hold on to. The mouth stretches wide, the eyes squeeze shut, the cheeks rise and the head tilts forward. Every one of those changes pulls the face away from the calm, neutral shape in your photo, and the model has to invent the in-between frames.
When the source photo is weak, that invention goes wrong. A cropped, angled or already laughing picture gives the model less to anchor on, so mid-fit the person can look like a cousin of you rather than you. A laughing fit also moves fast, so a frame that looks fine alone can look odd in motion.
Which photo gives the best result?
Use one sharp photo of one adult facing the camera, with face and shoulders in the frame and a neutral, closed mouth. A calm front-facing picture gives the model the cleanest face to turn into a laugh. A photo that is already mid-laugh can stretch oddly.
Daylight or a bright room keeps skin tone and features steady through the fit. Skip hats and heavy or dark glasses: the video replaces the man's head completely, so these are often lost or half drawn. Avoid group photos too, because the clip has one main character and the model will pick one face or blend several.
How does a single photo become the full 10 second clip?
videosi uses ByteDance's Seedance 2.0 and 2.5 video models, and you pick the tier. A short motion reference supplies the sofa scene, the plate, the cut to the railing and the timing of the laughter. Your photo is the only source for the face, hair, build and clothes, in both scenes.
The instruction built into the page tells the model to swap the man in the reference for the person in your photo and keep everything else the same. That strict replace wording matters: without it a model tends to keep the original man and only borrow your clothes.
Why do many of the original clips carry a Kling AI mark?
Many of the clips in the wave show a small Kling AI mark in a corner. That tells you they were made with an AI video tool and exported with its mark intact. The wave is largely made of AI clips, and the mark is simply what the tool leaves behind.
Our clip does not carry that mark. videosi makes its own video from your photo, and the MP4 has no watermark. That does not remove the need to say it is AI. A clean file is easier to mistake for real footage, so the label matters more, not less.
How is this different from a filter or a CapCut template?
No single filter or template is confirmed for this trend. A face filter changes your live camera feed or lays an effect over your own footage. This video is generated: a model builds new frames of a man on a sofa, a cut, and a second scene, and gives him your face.
That is why it needs one good photo and a few minutes rather than a tap. It is also why you can make it without ever filming yourself. The trade-off is that generation can drift from your likeness, which a filter on your real face does not.
If you want the quick path, upload a photo on the ACP Pradyuman AI page, choose a format and press Generate.
Should you label the video as AI, and how?
Yes. Say it plainly in the caption, for example 'AI-generated video' or '#AI', and use the platform's own AI label setting where it exists. TikTok, Instagram and YouTube all ask creators to disclose realistic synthetic media, and a clear label lowers the risk of a post being flagged as misleading.
The clip is a joke, but it still shows a face doing something it never did. A clear label tells viewers it is a gag and protects you if someone shares it out of context. Put the label in the caption itself, not only in a comment, so it travels with the video.
Also check the platform's rules for the sound. The audio is comedy audio, not a licensed song.
Is it okay to make this with a real person's likeness?
Only with their permission. Use your own photo, or the photo of a friend who has agreed to be in the clip. Do not use a photo of someone who has not said yes, and do not use a stranger, a colleague or a public figure without consent.
Do not post a result as if the real person said or did something. The format is built for self-parody, where you are the stern one who cracks. It is not built for mocking someone else or making them look foolish without their knowledge.
videosi asks for photos you have the right to use, supports adults only, and does not make clips that claim a real person did something. Pets are not supported on this page.
What does one video cost?
A 10 second video at Standard 720p is 10 credits, and the higher tiers cost more. The Single pack of 10 credits is $8.90, or $5.99 in India, so one pack covers one Standard video. There is no free tier for this page.
A failed, timed-out or blocked render returns its credits automatically. Because a laughing fit can drift, it is worth spending a minute on the photo before you press Generate rather than running several weak tries.
For more formats like this one, see what is trending this week.



