What aura farming Korean AI is, and what it is not
“Aura” is internet slang for the cool presence someone gives off, and “farming” is borrowed from games, where you grind for points. Aura farming is doing something calm and effortless on purpose to collect those points. Put “Korean AI” after it and people mean one video format: a K-drama style main character entrance into a Korean classroom, made from your own photo.
| Search | What it means |
|---|---|
| Aura farming Korean AI (this page) | A Korean school-drama classroom: classmates dancing on the desks, a teacher glaring from the green board, a transfer student who walks in calm and gets the red and blue neon close-ups with sunglasses pushed up. People put themselves in as the student, from a photo. TikTok tutorials from 9–10 October 2026. |
| Aura farming (the slang) | Acting effortlessly cool for “aura points”; it spread through TikTok edits in 2025. A word, not a video template. |
| Korean natural farming, AURA brands | Unrelated: an agriculture method and Korean product brands that share the words. Nothing to do with the AI trend. |
The classroom version is often described as inspired by the Korean drama Study Group. The clip this page follows has AI-generated actors, not the drama’s cast, and keeps the trend’s song.
Korean aura farming AI vs editing only the face
| Aura farming Korean AI | Face-only edit | |
|---|---|---|
| Goal | You become the main character of one K-drama scene | Your face on someone else’s body |
| Input | Two photos (student and teacher) plus the classroom reference clip | A target video and a face image |
| What changes | The whole person: face, hair, build and clothes from each photo | Only the face; the original hair and T-shirt stay |
| What stays | The entrance, the look-back, the neon cuts and the song | Everything except the face |
| Typical problem | Stiff motion when a tool invents it (solved by the reference clip) | A mismatched head, the original look leaking back |
Template, prompt, video edit or reference clip: four ways to make it
The scene is about timing: the walk-in, the glance back, the cut to neon. How you make it decides whether that timing survives.
| Way | Who moves the person | Does your face hold? | The song | Good for |
|---|---|---|---|---|
| CapCut-style template | Nobody: your own photos and clips in preset cuts | Yes, it is your footage | Included | A fast, free meme edit; you are not in the classroom |
| Text or image-to-video prompt | The model guesses the walk and the glance | Often drifts; motion looks stiff | Add it yourself | A loose “main character” clip, not this scene |
| Chat video edit (“replace the student with me”) | The clip, if the tool really reads it | Varies; the original face can leak back | Often lost | Trying it once with a tool you already use |
| Reference clip + photo (this page) | The scene’s own clip, cut by cut | Yes: the photo is the only source of the person | Restored from the clip | Being the transfer student, with a friend as the teacher, in the actual scene |
Short version: templates are fastest, prompts are the most flexible, and a reference clip is the only one that copies the scene’s motion instead of imitating it.
Why Korean aura farming AI videos look fake, and the fix for each
| What you see | Why it happens | What to change |
|---|---|---|
| The classroom entrance looks robotic or floaty | The model invented the motion from a still photo | Drive it with the scene’s clip (the panel above) instead of a prompt |
| Your face changes in the neon close-ups | Colored light and fast cuts, plus a small or angled photo | A sharp front-facing photo and 720p or above |
| You look like a different, “prettier” person | Beauty filters in the photo, or prompts asking for a perfect face | An unfiltered photo; describe nothing about the face |
| Sunglasses melt into your face | Your photo already has sunglasses, so the model merges two pairs | Upload a photo without sunglasses; the scene adds its own |
| The original student’s T-shirt or hair comes back | The tool only swapped the face | Use a photo that shows your top; here the whole person comes from the photo |
| You and your friend blend into one face | Two similar photos, or a group photo for each slot | One person per photo, different framing, both faces clearly lit |
| Blurry after posting | 480p plus a second export at a low bitrate | 720p or 1080p; upload the downloaded MP4 without re-exporting |
Pick a photo that holds up in the neon close-ups
The last third of the scene is close-ups in red and blue light. That is where a weak photo shows.
| Photo | Result |
|---|---|
| Front-facing, upper body, even light, one adult | Best: you stay recognisable through every close-up |
| Face-only selfie | Works; the top is invented to match, and the face drifts a little more |
| Old or low-resolution photo | Fine if the face is sharp at phone size; blurry scans lose the likeness |
| Group photo | Crop to one person first, or the model may pick the wrong one |
| Side profile, sunglasses, cap or beauty filter | The identity drifts; use a plain front-facing photo |
| A child, a celebrity, a K-pop idol or a stranger | Not allowed: adults who agreed only |
The K-drama classroom shot by shot: who plays what
The scene runs 30 seconds in three beats: the class before you arrive, the main character entrance, and the reaction close-ups. Your photo is the transfer student in every shot; your friend’s photo is the teacher in every teacher shot.
| Time | What happens | Who |
|---|---|---|
| 0–7.5 s | Classmates climb onto the desks and dance; the teacher talks, then turns and glares | Teacher |
| 7.5–12.7 s | The transfer student walks in at the back, stops and looks back over the shoulder | Student |
| 12.7–18.5 s | The student sits at a back desk and smiles over a notebook while the teacher keeps staring | Student, teacher |
| 18.5–27 s | Red and blue neon close-ups: sunglasses pushed up, a grin on folded arms, a half-turn, the hood shot with sunglasses on | Student |
| 27–30 s | Arms crossed at the desk, smiling, held to the end | Student |
Seedance 2.5 makes all 30 seconds in one run, so there is no join in the middle. For generation the original student and teacher are faceless stand-ins, so nothing of the original actors’ faces, hair or clothes carries into your video; only your two photos decide who they are.
Posting it on TikTok, Reels and Shorts
- Format. The MP4 is square (1:1), like the scene itself, with the song under it. Square plays as-is in Instagram and X feeds; TikTok, Reels and Shorts show it with bars above and below unless you place it on a 9:16 canvas in your editor. Pick 720p or 1080p if the clip is going straight to TikTok or Reels, and upload the downloaded file without re-exporting.
- Length. The 30-second scene tells the whole entrance; trim the last few seconds in the app if you want a tighter loop.
- The sound. The trend’s song comes with the clip. If you want the original sound linked on TikTok, add it from the app’s sound library at low volume instead of replacing the audio.
- Captions. Add text in the app, not in the video: “POV: the transfer student just walked in”, “Episode 1: the new kid”, “Aura +1000”.
- Label it. Turn on the platform’s AI-generated label; the trend is obviously made with AI, and saying so avoids takedowns.
Aura farming Korean AI price: one video and a batch of 20
You pay per video in credits, and the exact price is on the Generate button before you start. Every video is the whole 30-second scene on Seedance 2.5 with both photos, so the price only depends on the resolution: 233 credits at 480p, 523 at 720p and 928 at 1080p.
| 模型 | 30 秒 | 20 × 30 秒 |
|---|---|---|
| Seedance 2.5 · 480p | 233 | 4,660 |
| Seedance 2.5 · 720p | 523 | 10,460 |
| Seedance 2.5 · 1080p | 928 | 18,560 |
On the Pro monthly plan (780 credits for $29), one credit costs about $0.037: one video is about $8.66 at 480p and $19.44 at 720p, and 20 videos at 480p come to about $173.26.
| Free | Signing up and uploading photos. New accounts get 3 free credits, which is not enough for a video: the cheapest one costs 233. There is no unlimited free version. |
|---|---|
| Smallest purchase | $5 Trial Pack = 24 credits, not enough for one video here (233 credits at 480p); a monthly plan or a larger pack covers it |
| Watermark | Videos made before your first purchase carry a small moving genjutsu-ai.app watermark; any purchase unlocks the clean originals, including videos you already made. |
| Generation time | Usually a few minutes. Seedance 2.5 reviews the photos and the scene first. You can leave the page; the video stays in your history. |
| Failed or rejected | Credits are refunded automatically |
Have your own entrance clip? Recast it instead
The panel above is for the Korean classroom scene. If you already filmed something with the same energy, such as a slow walk down a school hallway, a café door entrance or your brand mascot striking a pose, open the full editor and upload it as the source (5 s to 5 minutes of MP4 or MOV). Pick Character to put another person in, Outfit to restyle what they wear or Background to move them somewhere else, and the clip’s own motion and camera stay. The motion transfer guide explains how the reference drives the result.
A copy-ready K-drama transfer student prompt for other tools
If you use a tool that accepts a reference video and images, attach the classroom clip as the video, your photo as the first image and your friend’s as the second, then paste this. Rename “reference video” and “reference image” to whatever the tool calls them (for example @Video1, @Image1 and @Image2).
Use the reference video only for its shots, camera moves, timing and acting. Recast the transfer student who walks into the Korean classroom, looks back over the shoulder, sits at the back and gets the red and blue neon close-ups as the person in the first reference image, in every shot. Recast the teacher at the green board as the person in the second reference image, in every teacher shot. Each image decides how that person looks everywhere: face, hair, skin tone, build and the clothes they wear. If an image stops at the shoulders, continue the same top naturally. Nothing of the original student or teacher is left. Never blend the two people. The dark sunglasses are the student's only prop: pushed up on the head in the class shots, over the eyes only in the hood shot. Keep the classmates dancing on the desks, the neon light and every cut at normal speed. Two recast people, natural hands, the same square frame as the reference, no captions, logos or watermark.
Keep the clip to the length your tool allows; many reference modes stop at 15 seconds, so send the scene in two halves and join them.
different person, changed identity, beautified or slimmed face, anime face, plastic skin, two pairs of sunglasses, extra fingers, warped desks, flickering face in the neon shots, slow motion, captions or watermark
Keeps your identity instead of a beautified stranger. Leave out words like “perfect face” or “handsome” in the main prompt for the same reason.
A text-only prompt can describe a cool entrance, but it will not reproduce this scene’s cuts or the song. For more on writing instructions for any clip, see the Genjutsu AI prompt guide.
Aura farming Korean AI examples
The same classroom scene with three different students and teachers.
Brands, the song and real people
- Brand and client work. The K-drama classroom look is a trend, not a trademark, but the song under the video is not licensed by buying credits. For ads, swap in a commercial track or license the original, and follow each platform’s AI-content labels.
- Real people. Only use your own photo or photos of adults who agreed to be in the video. No celebrities, K-drama actors, idols, strangers or minors; you confirm this before each generation.
- Your photos. They stay private inputs on your account and are never shown on the site. See the privacy policy for retention and deletion.
Genjutsu AI is an independent tool, not affiliated with any drama, the song’s rights holders, Higgsfield or CapCut.
More trend templates: the zombie AI trend, the two-person Raindance pier duet and the Love Is Gonna Save Us dance. For two people in any clip, see duo swap; for free credits and the watermark, see Genjutsu AI free.
最近更新: 2026-10-15