The couple expecting a baby
A tender moment for two during the pregnancy: you imagine the baby, laugh at which features came from whom, and keep the picture as a keepsake.
AI baby generator: imagine your future baby's face
Two photos, one of each parent, and the AI imagines your child at one and at six with a mix of your features. Just for fun, no account needed to see the workflow.
Real workflow: Frank's selfie and the portrait of an invented woman go into two Nano Banana Pro nodes, which imagine their child at one and then at six.
At a glance
- You provide
- 2 photos · a text or a brief
- What the workflow produces
- 2 images in 3:4
- Canvas structure
- 7 nodes wired by 6 connections
- Models wired in
- Nano Banana Pro
- Cost of one full run
- about 38 credits, roughly $0.46
- Account
- The demo is visible without an account; generating requires a free account, trial credits included
"What will our baby look like?" Every couple asks it, laughing. This workflow takes a photo of each of you and imagines your child twice: at one, on a blanket, then at six, in a park with a missing tooth.
The hard part is the mix. Give a model two faces and it happily copies the sharper one. The prompt splits the features, the eyes and smile of one, the hair and brow of the other, a skin tone in between, and both ages start from the same photos so they stay the same child.
It is a dreamed-up picture, not a prediction: genetics can't be read off a selfie. To watch the child grow up, chain it with age progression; to put them next to you, the family photo. Both pictures cost thirty-eight credits.
Actual renders from the demo
These are the files of the workflow above, exactly as the models produced them, unretouched. The starting point is shown when there is one.
Starting point
Render
Who it is for
Couples expecting a child or thinking about one, and anyone curious to see two faces blended.
Which files to provide
A front-facing photo of each parent, plus the age and scene wanted for the child.
How it works
- 1
One photo of each parent
Two front-facing, well-lit photos, one person per photo, no sunglasses. The sharper the face, the more of its features reach the child.
- 2
Say which feature comes from whom
The eyes and smile of one, the hair and brow of the other, a skin tone in between. The model blends instead of copying.
- 3
Pick the age and the scene
One on a blanket, six in a park, ten at the beach: one age and one setting per node, the same features everywhere.
- 4
Compare the ages
Both pictures start from the same photos, so you recognise the same child from one age to the next. Add a node for a third age.
Variations and use cases
The same canvas, tuned for different needs: each variation is two or three lines of prompt away.
Sharing the pregnancy news
The imagined child becomes the playful touch of a message to family and friends before the scan, with a note that it is only a dream.
A game between friends
Two curious friends, a wedding coming up, a party: you run the workflow and compare the imagined children, without taking any of it seriously.
A character in a story
An author imagines the child of two characters from a novel at several ages, then uses the same character for the rest of the story.
The child grown up
The six year old then goes through age progression to see the teenager and the adult they might become.
The dream family photo
Parents and imagined child brought together in one picture with the group photo, for a gift or a card.
Common mistakes
Not splitting the features
Without saying which feature comes from whom, the model copies the parent whose photo is sharpest. The eyes of one, the hair of the other.
Side-on or blurry photos
A face you can barely see passes nothing on, and the child looks like one parent only. Two sharp, front-facing, well-lit photos.
Leaving the age vague
"A child" lets the model pick anything from two to ten, and the features drift. Give an age and a detail that shows it.
Presenting it as a prediction
The picture is a dream, it says nothing about the real baby. Shared without saying so, it can disappoint or mislead: say it is imagined.
Recommended models
Nano Banana Pro
GPT Image 2.5 Sunburst
Seedream 5 Pro
Flux KontextFrequently asked questions
Is this a prediction of my real baby?
No. It is a picture imagined from your two faces, for fun. Real genetics can't be read from a photo.
Can I choose a girl or a boy?
Yes: write it in each node's prompt. Without it, the model picks on its own, as in the demo.
Do the photos need to be recent?
Ideally, front-facing and sharp. An old blurry or side-on photo passes its features on poorly.
How much do the two pictures cost?
Thirty-eight credits, nineteen per picture on Nano Banana Pro. A third age costs nineteen credits more.
Why Nano Banana Pro?
Because it holds several reference faces at once and blends their features, where other models fall back on a single face.
Is it the same child in both pictures?
In the demo, yes: same curls, same smile, same skin tone at one and at six, because both nodes get the same photos and the same features.
What output format?
3:4, the shape of a family photo. Change it in the node for a square to share.
Are my photos kept?
They stay in your space, like all your media, and are only used for your own generations.
AI baby generator: imagine your future baby's face
Use this workflowFree account, no card required. The workflow opens pre-filled in your canvas.