Short answer: open the Talk to Me Doja AI Video Generator, pick You, Duo or Pet, upload one photo per person or pet, and press Generate. A few minutes later you have a 13-second vertical MP4: the cool build-up, both palms pressed to the cheeks on the beat drop, and the shy side-to-side sway. Post it with the sped-up "Like That" sound from the TikTok library and the caption "talk to me doja". No app, no editing, no prompt.
What you need to make a Talk to Me Doja AI video
- One photo per person or pet. For You: one adult, waist up, face to the camera, arms in the picture. For Duo: two photos, one adult each. For Pet: the face and both front paws in view.
- A few minutes. The video renders in the cloud; you can leave the page and find the clip in your library when it is done.
- Credits for one clip. The exact price is on the Generate button before you spend anything, and you can buy credits without creating an account.
You do not need the anime GIF, the song file or any editing app. The caption and the GIF sticker are added when you post, and the sound comes from the TikTok or Reels library.
Step 1: pick You, Duo or Pet
The generator has three versions of the same sway:
- You — the classic take. One photo and you do the whole build-up and sway yourself. The more serious you look, the funnier the switch.
- Duo — two photos, one person in each. The two of you stand side by side and do the cheek-cup sway in sync; photo 1 always stands on the left.
- Pet — one photo of your cat or dog. It sits up and presses its front paws to its cheeks, bashful like the anime girl.
All three are 13 seconds, vertical 9:16, made for TikTok, Reels and Shorts.
Step 2: upload the photo
Upload your photo (or photos, for Duo). A few things make a visible difference:
- Sharp and front-facing. The face is the whole joke, so pick a photo where it is big, in focus and turned to the camera, in good light.
- Waist up, arms in view. The clip is a medium shot and your hands go up to your cheeks, so the AI needs your arms.
- A straight face helps. A deadpan or serious photo makes the shy sway land harder.
- No sunglasses, masks or face-hiding hats, and no heavy beauty filters — they confuse the identity.
- Only yourself, an adult who agreed, or your own pet. No children, celebrities or strangers.
Step 3: generate
Press Generate. The button shows the price first; you only sign in or buy credits at this step. The clip takes a few minutes, and you can leave the page: it waits in your library when it's done. If generation fails, the credits come back automatically.
The result is one continuous take in a bright room: six seconds of playing it cool, the palms-to-cheeks drop at second 6, and the bashful sway until the end. It carries its own upbeat beat, so it already works on its own.
Step 4: post it with the Talk to Me Doja sound
The trend lives on its sound, and your clip is timed for it:
- Download the MP4 to your phone.
- In TikTok or Reels, add the sped-up "Like That" from the app's own sound library. It starts at second 0, so the beat drop lands exactly on the moment your hands hit your cheeks. Using the platform library also means the song is licensed for your post.
- Add the caption "talk to me doja" in the middle of the screen. Many creators also park the anime GIF in a corner with the sticker tool.
- Hashtags: #talktomedoja, #speaktomedoja, #urarashiraishi. Keep the platform's AI label on.
What can go wrong, and how do you fix it?
| Problem | Fix |
|---|---|
| The face looks less like you | Use a sharper, front-facing photo in good light and run it again. |
| The hands on the cheeks blur or gain fingers | Hands on faces are the hardest moment for any video model. Each run comes out different — a second run usually fixes it. |
| The sway starts a little early or late | The clip follows the beat, not the caption. If the drop does not land where you want, run it again. |
| Duo: the two people swap sides | Photo 1 always stands on the left. Check the order before you press Generate. |
| Pet: a paw looks like a hand | Use a photo where both front paws are clearly in view and run it again. |
| The photo is refused | Some photos are refused by the safety check. The credits come back; try another photo. |
How much does it cost?
You pay per clip in credits, and the exact price is on the Generate button before you start. The default is 720p; 480p under Advanced is the cheaper take. Credits work on every app on Vuddex, credit packs are a one-time payment with no subscription, and a clip that fails to generate is refunded.
More about the trend
Wondering where the shy anime girl comes from, which episode the GIF is from and why the caption says "Doja"? Read Talk to Me Doja: the anime girl sway meme explained — the Yamada-kun scene, the Doja Cat song and the week the trend took over TikTok.