Skip to content
Lover Snap

AI photos of deceased loved ones: a 2026 guide for grieving families

A practical guide to using AI photography to honor deceased family members in 2026 — how the technology works, which apps to consider, what to expect, and the ethical questions worth thinking through first.

By Jiuhong Deng · · Updated

If you’re searching for AI photos of a deceased loved one, you’re not alone. In 2026 this is one of the fastest-growing search categories in consumer AI — and one of the most emotionally loaded. This guide explains how the technology works, what tools exist, what to expect, and the questions worth thinking through before you start.

What “AI photos of a deceased loved one” actually means in 2026

The phrase covers four different technical jobs, and they’re often confused with each other:

  1. Photo restoration. Taking a damaged, blurry, or low-resolution photograph of someone who has died and improving it. Best tool: Remini, or professional services. No new image content is generated.
  2. Single-photo animation. Taking one existing photo and adding short looping motion — a head turn, a smile, a blink. Best known: MyHeritage Deep Nostalgia.
  3. New photo generation. Training an AI on multiple reference photos to create photographs of the loved one in scenes that never happened — a wedding, a holiday, a quiet moment. Best tool: Lover Snap.
  4. Conversational legacy avatars. A chatbot trained on recordings of the person while they were alive, allowing surviving family to interact with a recording-driven simulation. Best known: Hereafter AI.

The four are technically different and emotionally different. Pick the job before you pick the tool.

How AI photo generation works for memorial use

Most AI memorial photo tools in 2026 use one of two underlying image engines: Black Forest Labs’ Flux (which Lover Snap uses) or Stability AI’s SDXL/Stable Diffusion 3 family. The Flux model produces notably more photorealistic outputs and is becoming the default for memorial-grade work.

The workflow is roughly the same across tools:

  1. Upload reference photos of the deceased person — typically 3 to 16.
  2. The system trains a private model that learns their face and features (~30 minutes).
  3. You generate new photographs in scenes you choose (~23 seconds each).
  4. Keep what feels right, delete what doesn’t.

What 2026 looks like for the field

A few specific shifts have happened since AI memorial photography emerged in 2023-2024:

What to expect emotionally

Three things worth knowing before you start.

Picking the right tool

If you want…Use
To repair a single damaged photoRemini, or a professional photo restorer
Subtle motion on one existing photoMyHeritage Deep Nostalgia
New photographs of you with the deceasedLover Snap
To converse with a recording-driven avatarHereafter AI (requires pre-death recordings)

For most bereaved families looking for new photographs they didn’t get to take during the loved one’s lifetime, Lover Snap is the answer. For restoration, Remini. The other categories serve specific needs that not everyone has.

A note on grief support

AI memorial photography is a tool, not therapy. If you’re in acute grief, please consider working with a grief counselor or peer-support group in parallel. We maintain a list of resources at /memorial/grief-resources. The most useful framing we’ve heard from grief researchers: AI memorial photos can be part of how you honor memory, but they should not be the only way you process loss.

Further reading