КАДР

How to make a video greeting with AI: animate a photo, add speech and music

Alen founder of KADR Updated 2026-10-11 4 min read

An AI video greeting is made from a single photo in three steps: prepare a festive shot (for example, a New Year portrait in Nano Banana Pro), animate it with a video model and add sound: speech, music or holiday ambience. Veo 3.1 and Kling v3 generate video together with sound, so the character can say a short greeting themselves. A 6-second clip in 1080p with sound costs from 84 rubles on Veo 3.1 Lite to 168 rubles on Kling v3 (KADR data, October 2026). Below: step by step, ready prompts for New Year and birthdays, and what to do if the model refuses.

Three steps

  1. Festive shot. Turn an ordinary photo into a New Year or holiday portrait.
  2. Animation. The video model adds movement: a smile, falling snow, twinkling garlands.
  3. Sound. Speech, music or holiday ambience. Models with sound generate it together with the video.

Step 1. Festive shot

If you already have a suitable photo, skip this step. If not, upload an ordinary portrait to Nano Banana Pro and describe the festive look:

Keep the face and features of the person in the photo unchanged. Cosy New Year portrait at home, red knitted sweater, a mug of cocoa, a Christmas tree with warm lights in the background. Format 9:16.

Choose the format for the platform right away: 9:16 for messengers and stories, 16:9 for a screen. 12 New Year ideas with prompts: New Year AI photoshoot.

Step 2. Animation

Upload the shot to a video model in «from photo» mode (image-to-video). On KADR there is the «Animate photo» mode for this.

Which model to choose:

Model When to use 6 s, 1080p, with sound
Veo 3.1 Lite simple animation, atmosphere, inexpensive ~84 rub.
Veo 3.1 Fast better light and detail, speech ~126 rub.
Kling v3 a real person in close-up, facial expressions, speech ~168 rub.
Veo 3.1 maximum quality, speech ~417 rub.

Prices on KADR, October 2026. This is not a public offer.

The animation rule: the less movement, the better the face holds. A smile, a head turn, a wave of the hand work. Dancing and running in close-up often break the likeness.

Step 3. Sound and speech

Models with sound generate it from the description in the prompt. There are three options.

Ambience: «crackling fireplace, quiet New Year melody, clinking glasses».

Speech: a short phrase in quotation marks, and the character will say it: «says to the camera: «Happy New Year, Mom!»». The line can be in Russian too («С Новым годом, мама!»), but in Russian the lips match the speech worse than in English, so keep the phrase short, 3 to 6 words.

Music on top: generate the video without speech and add a track and text in a video editor. This makes a long greeting easier.

More about talking characters: AI avatar and talking head.

Ready prompts

New Year, smile and snow:

The person in the photo smiles softly and waves at the camera, snow falls outside the window, the Christmas tree garlands twinkle. Static camera. Sound: quiet New Year melody, crackling fireplace.

New Year with a greeting:

The person in the photo looks into the camera, smiles and says: «Happy New Year! Wishing you happiness!». Warm garland light, slow camera push-in. Sound: quiet background music.

The spoken line can be written in Russian if the greeting is for a Russian-speaking audience.

Birthday:

The person in the photo blows out the candles on a cake, confetti flies around, everyone laughs. Slow push-in. Sound: applause, cheerful shouts.

Old family photo:

The people in the photo slightly turn their heads and smile, light natural movement, the camera is still. Sound: quiet warm melody.

Not sure about the wording? The «Improve» button on KADR will rewrite the prompt and suggest a model.

If the model refuses

A refusal is almost always about moderation: a realistic face, a child in the frame, a famous person. Seedance filters real faces more strictly than others, Kling is more relaxed about them. Change the model before rewriting the prompt. And a rule: make greetings with your own photos or photos of people who do not mind. You cannot use celebrities’ voices and faces: AI voice cloning.

How much a whole greeting costs

Option What it includes Approx.
Simple the photo as is + Veo 3.1 Lite, 6 s ~84 rub.
With a festive shot Nano Banana Pro + Veo 3.1 Fast, 6 s ~150 rub.
With a real person speaking Nano Banana Pro + Kling v3, 6 s ~190 rub.
Long, 3 scenes 3 Nano Banana Pro shots + 3 Veo 3.1 Fast clips ~455 rub.

To account for failed attempts, multiply by 1.5 to 2. How to animate a photo in detail: how to animate a photo with AI.

In short

An AI video greeting is made from a single photo: a festive shot in Nano Banana Pro, animation in Kling v3 or Veo 3.1, sound and a short line right in the prompt. A 6-second clip with sound costs from 84 rubles. Keep the movement minimal and the phrase short, and the face and speech will look natural.

Frequently asked questions

How do I make a video greeting with AI?
Take a photo, turn it into a festive shot in Nano Banana Pro if needed, then animate it with a video model such as Kling v3 or Veo 3.1 and describe the movement and sound in the prompt. You can put a short line in quotation marks, and the model will say it.
Can AI make the person in the photo say a greeting?
Yes. Veo 3.1 and Kling v3 generate speech together with the video: write a short phrase in quotation marks, for example «Happy New Year, Mom!» (it can also be in Russian: «С Новым годом, мама!»). In Russian the lips match the speech worse than in English, so keep the phrase short.
How much does an AI video greeting cost?
On KADR, 6 seconds in 1080p with sound: Veo 3.1 Lite about 84 rubles, Veo 3.1 Fast about 126 rubles, Kling v3 about 168 rubles. Plus about 25 rubles if you first make a festive portrait in Nano Banana Pro (October 2026).
Which AI is best for a greeting with a real person in the photo?
Kling v3: it keeps the face and expressions well and is more relaxed about photos of real people. Seedance may refuse a request with a realistic face.
Can a greeting be longer than 6 seconds?
Yes, assemble it from 2 to 4 short clips in any video editor and add music. Kling v3 makes up to 15 seconds at a time, Veo 3.1 up to 8 seconds.
Can I make a greeting in a celebrity's voice?
No. You cannot use a famous person's voice and likeness without their consent. Send greetings in your own name or with a voice the model generates.

A greeting from a single photo

On KADR: a festive portrait on Nano Banana Pro, animation on Kling v3 or Veo 3.1 with sound and speech. The «Animate photo» mode and the «Improve» button will help with the prompt. Pay in rubles, tokens never expire.

Make a greeting
Alen
Founder of KADR. Video greetings are one of the most common tasks among our users before the holidays.