Two photos in · One future baby out · Free credits to start
Upload a photo of each parent and get back a baby who could plausibly belong to both — his brow, her smile, blended into one small face. Made for the question every couple eventually googles at 1am.
A real run from this page's model: the two generated parent portraits on one side, the baby the model blended from them on the other. Drag the divider and trace where each feature came from — the curls, the eye set, the smile. Every output is AI-generated; your baby will look different because your faces do.
Drag to compare before and after
It is a two-photo image tool: you give it one clear photo of each parent, and it generates a new photorealistic baby face that inherits believable features from both. It answers the oldest what-if in every relationship with a picture instead of a guess.
Three kinds of people type this search. Couples — the biggest group by far — who want tonight's laugh or a soft way to start a bigger conversation. Friends making a wildly premature wedding-toast slide. And expecting parents who want a face to daydream about before the scan photos sharpen. None of them want genetics; all of them want the picture.
The pair-to-baby example on this page is our own run: we generated two parent portraits, fed both to this page's model, and asked for a one-year-old who blends them. Look at the result against the sources — the curl pattern and warm undertone from one parent, the eye set and cheek line from the other. No retouching after the fact; that frame is what came back.
Honesty about what this is: entertainment, not embryology. The model blends what faces look like, not chromosomes — it cannot know whose nose 'wins'. Run it five times and you get five plausible siblings, which is the correct behaviour for a question that genetics itself answers with a dice roll.
Creative engine
Upload both photos in one request, describe the age you want, and generate — the blend comes back in under a minute.
Both parents go into one generation request — the result inherits visible features from each side, and you can trace which came from whom.
Newborn, one-year-old, five on a beach — the age and the scene are words you choose, not presets you scroll.
Two uploads and one sentence — the phrasing sets the age and the mood.
Front-facing, decent light, one face per photo. The blend inherits what it can see, so sunglasses hide exactly the feature you most want passed down.
"A one-year-old in soft nursery light" reads very differently from "a five-year-old on a beach". Age, light and framing are yours to set in a sentence.
Treat the first result as one possible child, not the verdict. Rerunning deals a new genetic hand each time — most couples keep the third or fourth.
The same picture doing three different jobs.
The question comes up on its own; now there is a picture instead of a shrug. Fair warning: the result gets saved, sent to mothers, and argued over affectionately — and the aging filter answers the other late-night what-if.
Two friends' photos, one baby, instant chaos. It is the rare AI gag that lands harder the better the blend actually is.
Between the announcement and the first ultrasound photo worth framing, a generated face gives the daydream something to hold onto.
What people ask before uploading two faces.
Both photos go into a single generation request, and the instruction names what to take from each. You can verify it yourself the way our demo does: generate once, then compare the baby's features against each parent — the inheritance is visible, not implied. A random-baby site cannot survive that comparison.
Yes — both live in the sentence, not in a settings panel. "A two-year-old girl with her father's eyes" is a complete, valid instruction. If you leave it open, the model decides, and rerunning re-decides.
No. It blends appearance, not DNA — real inheritance runs through chromosomes the model never sees. Treat it as a very good artist's impression of a plausible child, which is also why every rerun differs: real genetics would too.
A blend runs on credits, and new accounts start with a free grant — enough to generate and rerun a few times before deciding whether it is worth anything. The estimate shows in the generator before you commit a run.
Uploaded over HTTPS to run the generation, stored so you can revisit results, never used for advertising or training showcases without consent. Delete them from your history and they are gone. Every output is AI-generated.
That is the honest limit of a two-photo blend: when one face is much sharper, larger or better-lit than the other, the model leans toward the photo it can read best. Matching the two photos in quality and framing is the single best fix — then rerun.
One credit pool covers Nano Banana images and Seedance and Veo video. The cost shows before every run, and failed jobs are not charged. Use a subscription for ongoing work, or a one-time pack when you just need to top up.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
One-time top-ups — buy extra credits any time you run low.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
Charges appear as “KAVEL AI” on your card statement. You can cancel any time from Settings → Billing; cancellation takes effect at the end of the period you already paid for.
Powered by
Use one balance across every supported image and video model — eleven of them today, listed below — and check the credit cost before each request.
Video models
Image models
Model credit guide
You're only charged for successful generations. The exact estimate in the generator varies by model, length, resolution, audio, and number of images.
| Type | Model | Credit cost |
|---|---|---|
| Video | Seedance 2.5 | The long-take tier, priced per second: 5s 480p ≈ 211 credits, 5s 720p ≈ 473. A full 30s take runs ≈ 1,260 at 480p and ≈ 2,835 at 720p. Supplying a reference clip lowers the per-second rate. |
| Video | Seedance 2.0 | 5s 720p image-to-video ≈ 188 credits; text-to-video ≈ 308 credits. Scales with resolution and length. |
| Video | Seedance 2 Fast | Faster and lower cost. 5s 720p text-to-video ≈ 248 credits. |
| Video | Seedance 2 Mini | The cheapest Seedance tier. 5s 720p ≈ 154 credits, 5s 480p ≈ 72 credits. |
| Video | Seedance 1.5 Pro | Audio doubles the rate. 5s 720p ≈ 27 credits silent, ≈ 53 with audio; 1080p ≈ 57 and ≈ 113. |
| Video | Veo 3.1 | Billed per video, not per second (Lite tier). About 45 credits at 720p and 53 at 1080p. |
| Video | Kling 3.0 | Audio raises the rate. 5s 720p ≈ 105 credits silent, ≈ 150 with audio; 1080p ≈ 135 and ≈ 203. |
| Video | MiniMax H3 | Fixed 2K, no resolution ladder. Priced per second — a 5s clip ≈ 158 credits. |
| Image | Nano Banana 2 | Generate or edit from text and images. About 8 credits per 1K image, 12 at 2K, 18 at 4K. |
| Image | Nano Banana Pro | Consistent run times across generations. About 12 credits per 1K or 2K image, 21 at 4K. |
| Image | Nano Banana 2 Lite | Faster, simpler variant. A flat 8 credits per image at every resolution. Start here to test. |
| Image | GPT Image 2 | The lowest-cost image model here. About 3 credits per 1K image, 6 at 2K, 12 at 4K. |
| Image | Seedream 5.0 Lite | Flat pricing — about 8 credits per image whether you generate or edit. |
Tools that answer the neighbouring what-ifs.
Two photos, one sentence, one small face that belongs to both of you — and if the result deserves a frame, [stage the wedding photo](/image/ai-wedding-photo-generator) or [see yourselves decades on](/image/ai-aging-filter) next.
A real edit made with this tool — drag your own photo in on the left, the prompt is already written, then press Generate.
