Skip to content
Lover Snap

Memorial · For a deceased parent

Photos with the mom or dad you miss.

The photographs people most often wish they had are photographs from milestones their parents didn't live to see. A wedding without a father. A first child's birthday without a grandmother. Lover Snap is built around exactly this — gently, with consent, and with a published ethics framework.

Reference photos that work well

  • Pick one era — your parent at the age you want to remember them
  • 3-16 photos minimum; more variety usually means a better character
  • Mix close-ups and full-body shots for natural duo-mode framing
  • Scan old prints at 300+ DPI; HEIC files need converting to JPG
  • Avoid hats, sunglasses, or anything that obscures the face

Scene ideas

  • · A walk in the family neighborhood — golden hour lighting
  • · Wedding photo — the photo that didn't get to be taken
  • · Holiday dinner at the family table — warm window light
  • · Sitting together on a beach or park bench
  • · Reading together, or working in the kitchen — anything quietly domestic

Common questions

What if my mom or dad died decades ago and I only have a few photos?
Three or four photos is enough to train a recognizable AI character — the AI can work with reference material from different decades. Variety helps more than count. If the photos are very old prints, scan them at high resolution first (300+ DPI).
Can the generated photo include other family members?
Yes. Train a separate AI character for each family member you want in the scene, then combine them in duo mode. Many users build the family photo over several months as they create characters for each parent or grandparent.
How do I handle photos of my parent at different ages?
Pick one era to train on per character. If you want photos of your parent in their thirties, train on photos from that decade only. Mixing eras tends to confuse the AI about which face features to favor.
Is this disrespectful?
We don't think so when it's done with care. We never generate photos of minors, never simulate speech, never claim the technology "brings someone back." See our full ethics framework for the boundaries we hold ourselves to.
Will my dad look exactly like the reference photos?
Resemblance scales with photo variety. About 75% of generated photos look unmistakably like the reference; 10% are exceptional; 15% have visible AI artifacts. Generate in burst mode (16 at once) and keep the 2-3 closest to your memory.

Ethics & dignity

Six commitments we hold ourselves to.

  1. 1

    Consent comes first

    We never generate AI photos of someone unless the person uploading them attests they have the right to do so. You are the steward of your loved one's memory.

  2. 2

    No minors

    We never generate photos of children or teenagers, living or deceased. This is a permanent product policy, not a default we can toggle.

  3. 3

    No simulated speech

    Lover Snap does not generate audio, deepfake video conversations, or any media that puts words in a deceased person's mouth. We make photographs, not impersonations.

  4. 4

    Private by default

    Reference photos and trained characters are stored encrypted, used only for your generations, never used to train shared models, and deleted within 30 days of account closure.

  5. 5

    Honest representation

    Every generated photo carries content credentials so it can be identified as AI-generated. We never market memorial photos as "real" or "indistinguishable."

  6. 6

    Grief support, not replacement

    Lover Snap is a tool for honoring memory. It is not therapy, not a substitute for grief counseling. We link to support resources for anyone who would benefit from them.

Have a concern or a story? [email protected]