Make an AI video greeting with your own people in it
An AI video greeting starts from a photo of your people and ends as a short clip where they move and you hear them: candles go out, glasses clink, someone cheers. You can pick one of the four video models we run for this - MiniMax H3 from MiniMax, Kling 3.0 from Kuaishou, Veo 3.1 from Google and Wan 3.0 from Alibaba - or just tell our agent in the chat whose day it is and let it ask the rest.
- from 380 credits
- 4-15 s
- With sound
Updated 29.09.2026
Four clips from four engines
Prompt
A kitchen birthday, candles out in one breath, then clapping and a big hug
Prompt
The family around the long garden table raise their glasses together and laugh, grandmother at the head of the table smiles and wipes a happy tear, leaves move in a light breeze, warm golden light, natural gentle camera push-in, sounds of clinking glasses, laughter and birdsong
Prompt
The family gathered around the birthday cake softly sings the last line of Happy Birthday, then Mum leans in and blows out the candles, everyone cheers and claps, candle smoke curls up, warm cosy kitchen light, slow cinematic push-in
Prompt
The elderly couple on the park bench lean their heads together and smile, he gently takes her hand, autumn leaves fall slowly around them, soft afternoon light, very steady camera, quiet ambience with birds
Photo in, clip out: three steps
- 1
Start with your people or a scene
Add their photos (JPG, PNG or WebP, up to 10 MB) or open Scenes and pick one of 20 video moments, such as "Make a wish, Mum". Your people go into the picture first; then the picture comes to life.
- 2
Say what should happen
Type one line to the agent in the chat, such as "A video greeting for Dad's 70th, the family raises a glass to him at the garden table", and it asks only what is still missing, one question at a time, with buttons. Or switch the composer to Standard, choose Video and set the length, the sound and the Model. Auto means MiniMax H3.
- 3
Watch it, then send it on the day
A scene video takes about 3-6 minutes: first the card, then the animation. If one thing is off, ask for that one change. Then say "send it to Dad on his birthday morning": the agent asks for the channel (email, SMS or a popular messenger) and the address, and shows one card to confirm.
Start from the moment: which engine for which clip
Start from the moment, not the model. For a birthday clip in the family chat, leave the picker on Auto: at 768p, MiniMax H3 is the cheapest per second of the four, always has sound and animates all 20 video scenes. The "Make a wish, Mum" clip above is H3 at 768p - candles out in one breath, then clapping and a hug.
Pick Kling 3.0 when the motion is the point: a whole table raising glasses at once while the camera moves in. It renders in 1080p and runs up to 15 seconds, the longest of the four. In the garden-table clip above, Grandma wipes a happy tear and you hear the clinking and the birds.
With sound on, Veo 3.1 costs the most per second, and it gives the cinema look, with sound that carries voices and singing - in the 6-second candles clip the family sings the end of Happy Birthday before Mum blows. It stops at 8 seconds and, from words alone, makes no square clip. Wan 3.0 Prime is for calm: a steady 5 or 10 second loop with one small gesture. If someone should say the greeting to the camera, make a talking photo card instead.
No prompt needed: the chat asks, you tap
You do not have to write a prompt. Type one line in the chat, such as "A video greeting for Mum's 60th", and the agent works through a short list, skipping anything you have already said: who it is for, what the day is, how old they are turning, who should be in the picture, a photo if the person has none saved, where it happens and what we should hear. Tap someone saved in People and it takes their birthday, photos and your notes from their card.
Every button is a shortcut, and you can always type instead. For the place and the sound, your own words usually win: "Outdoors, a picnic" gets a generic park, while "Grandma's garden, the long table under the fig tree" gets somewhere your family has actually sat. Under each question, "Your call" leaves that detail to the agent, and "No more questions" skips the rest and goes straight to the video.
At the end come a few chips with possible changes, a button to the scene catalogue and "Make the video from this description" with its price. The chips only change the plan; nothing paid starts until you press that button. Ask for one change at a time: "make it outdoors" is clear, while "outdoors, funnier, at night and with Grandpa" leaves the agent guessing which part you cared about. "Actually, let's make a song" switches the format and keeps what you already said. The chat never asks about the model and makes video on Auto, MiniMax H3; to pick Kling 3.0 or Veo 3.1, switch to Standard.
One action, one sound, one camera move
Five to ten seconds is one moment, not a story. The clips that work have one action with a clear end: candles go out, glasses meet, a door opens on people shouting surprise. Write that action plainly and say who does it. "Mum leans in and blows out the candles, everyone claps" beats "a beautiful birthday celebration", which gives the model nothing to animate.
Then name the sound: clinking glasses, a laugh, waves, birdsong, the last line of a song. The prompts behind the clips above all name something you can hear: the clapping, the clinking, the singing. Give the light one line and the camera one line - warm kitchen light, a slow push-in - and stop. Six actions in one prompt tend to give a crowded clip where nobody finishes anything.
Keep the cast small. A picture can hold up to six people, but faces stay most reliably themselves with four or fewer. For the whole family at one table, a still AI photo card is the better format, and a card can be brought to life later.
Scenes, timing and the evening before
Most video greetings here begin as a photo card: your people's photos go into the moment, then the card is animated and given sound. The Scenes catalogue does both steps for you - 36 ready moments, 20 of them video, from "One blanket, one sunset" to "First kiss as newlyweds". The scene pictures were drawn with Nano Banana Pro and animated with MiniMax H3. It works as a birthday video maker and just as well for a wedding, an anniversary or a new baby.
Plan for the clock, not the pixels. A scene video takes about 3-6 minutes, so make it the evening before rather than in the car outside the party. Then schedule it: the clip goes out by email, SMS or a popular messenger on the day.
If the first take is not quite right, change one thing - the action, the sound or the model - rather than rewriting everything. Each new take shows its price on the button first.
Seconds, sound and resolution across the four
Start on Auto (MiniMax H3) for value and the scenes. Move to Kling 3.0 for smooth motion or a 15-second clip, to Veo 3.1 for the cinema look with voices in the sound, and to Wan 3.0 for a calm, steady loop.
| Model | Strong at | Length | Quality | Sound | Price |
|---|---|---|---|---|---|
| MiniMax H3MiniMax | Best value | 5, 6, 8, 10 s | 768p, 2K | With sound | from 380 credits · 75 cr a second |
| Kling 3.0 ProKuaishou | Smooth motion | 5, 8, 10, 15 s | 1080p | With or without sound | from 700 credits · 210 cr a second |
| Veo 3.1Google | Cinema quality | 4, 6, 8 s | 720p, 1080p | With or without sound | from 1,000 credits · 500 cr a second |
| Wan 3.0 PrimeAlibaba | Steady loops | 5, 10 s | 720p, 1080p | With or without sound | from 880 credits · 175 cr a second |
Prices are in credits and are always shown on the button before you press it. Pricing
Written out: Mum's 60th, a golden wedding, Grandma's 80th
A video for my mum's 60th: she sits at the kitchen table, the family sings the last line of Happy Birthday, she blows out the candles in one breath, everyone claps and my brother hugs her. Warm evening light, slow push-in, we hear the singing and the clapping.
Grandma and Grandpa's 50th anniversary: the two of them at the head of a long garden table under string lights, the family stands and raises glasses to them, Grandma laughs and wipes a tear. Sounds of clinking glasses, laughter and birdsong.
For Grandma's 80th: she sits in her armchair by a sunny window, looks up from the handmade card the grandchildren sent, smiles and presses it to her chest, the curtain moves a little. Very steady camera, a quiet room, a clock ticking.
Keep exploring
MiniMax H3 is the AI video model from MiniMax: it turns a photo card, or a few lines about the…
OpenKling 3.0Kling 3.0 Pro is Kuaishou's video model, and of the three Kling models we run it is the only one…
OpenVeo 3.1Veo 3.1 is the full model of Google's Veo line and the top of the three Veo options in our…
OpenWanWan 3.0 Prime is Alibaba's video model, and on Wish Generator it is the one we point to for small,…
OpenTalking photo cardA talking photo card has someone you love say your greeting out loud - up to 15 seconds of speech,…
OpenVideo greetings from photos: what to know first
Add the photo, say what should happen, and choose a model. In the chat the agent asks a few questions one at a time; in Standard you pick Video, the length, the sound and the model, or leave it on Auto (MiniMax H3). Your people go into a card first, then the card is animated with sound. One picture holds up to six people, and faces stay most recognisable with four or fewer.
Yes. Tell the agent in the chat who the video is for and what the day is, and it asks the rest one question at a time, with buttons to tap. It writes the description for the video itself; you add the photos and press the button with the price on it. The chat makes video on Auto, MiniMax H3, so to choose another model, switch the composer to Standard.
Only for the video. Talking to the agent is free, and so are text greetings; you can even start before you sign in, and the app asks you to sign in when you press to make the video. The video costs credits, and every new account gets 1,000 free credits at sign-up. Of the four models on this page, a short clip on Auto costs the least. The price is on the button first, and a failed generation gives the credits back.
For most birthday videos, MiniMax H3: the best value, always with sound, and the engine of all 20 video scenes. Choose Kling 3.0 for smooth motion and up to 15 seconds, Veo 3.1 for voices and singing, Wan 3.0 for a calm loop. If Grandma should say the words herself, make a talking photo card.
Yes. MiniMax H3 always makes sound; Kling 3.0, Veo 3.1 and Wan 3.0 let you turn it on or off, and turning it off makes Kling 3.0 and Veo 3.1 cheaper. Write the sound into the prompt: clinking glasses, waves, a family singing. For a whole song with your own words, use the song generator (Google Lyria 3.5).
From 4 to 15 seconds, depending on the model: Veo 3.1 makes 4, 6 or 8 seconds; MiniMax H3 5, 6, 8 or 10; Wan 3.0 5 or 10; Kling 3.0 5, 8, 10 or 15. For a greeting, 5 to 8 seconds is usually enough - one moment with a clear end. Making it takes longer than watching it: a scene video takes about 3-6 minutes, so make it the evening before.
Yes. Schedule it by email, SMS or a popular messenger for the morning of the day, or tell the agent "send it to Mum on her birthday morning": it asks for whatever is missing, the channel and the address, then shows one card to confirm. Nothing goes out before you press. Save the date as a reminder too, and a week before next year's birthday you get a nudge.
A clip that arrives before the cake
Add their photo, write one moment and what it sounds like, and schedule the clip for the day.
Make a video greeting
