Live$AI / ML· June 2026

AI Pet Memorial

A gentle, empathy-driven AI platform that transforms a single pet photo into a custom-compiled storybook, high-quality portrait, and voice-narrated cinematic video tribute.

~/aipetmemorial ▌

// summary

Summary

AI Pet Memorial provides a gentle, supportive space to celebrate your pet's life. Instead of standard AI generations, it delivers a fully curated keepsake package—including custom-authored storybooks, comforting themed portraits, and dynamic, voice-narrated videos where your pet shares a message from a peaceful scene.

How it Works

The memorial creation flow is designed to be therapeutic, taking the user through a three-step wizard:

Share Memories: Pet parents gather their pet's favorite nicknames, foods, and quirks. Collecting these details acts as a comforting reflection on their journey together.
Choose a Background: Multiple peaceful templates are available to suit the pet's personality—ranging from the Rainbow Bridge and Heavenly Clouds to Beach Sunsets and cozy fireside living rooms.
Generate Keepsake: Laravel coordinates the multi-model AI pipeline, prompting OpenAI for custom storybooks, Eleven Labs for narration, and Hedra for avatar video generation. The final tribute is ready in minutes to download, print, or share.

// problem

Problem Solved

Losing a beloved pet is a deeply painful process. Standard photo editors are clinical and lack the emotional intelligence to help grieving owners construct a fitting tribute. Traditional video editing tools are too complex for grieving users. The goal was to build a secure, respectful, and simple space to turn a single photo into a complete, comforting keepsake package in minutes.

// features

Key Features

Step-by-step memorial compilation wizard to gather names, nicknames, and stories
OpenAI-driven storybook creation that crafts a custom narrative from pet memories
High-quality AI portrait generation set in serene environments like the Rainbow Bridge, Beach Sunset, or Heavenly Clouds
Eleven Labs integration for generating warm, sweet-toned voice narrations
Hedra AI video synthesis to create a talking, animated avatar of the user's pet
Private-by-default architecture deleting assets from active servers within 30 days

// architecture

Architecture Notes

A Laravel API backend that manages long-running queue workers for multi-model AI synthesis. Photos and text prompts are structured, filtered, and sent to OpenAI to compile a storybook and a voice-over script. Eleven Labs synthesizes the spoken narration, and Hedra processes the audio and reference image to render the final video. The front-end is written in React, leveraging Inertia.js for seamless routing and live background processing states.

// learnings

What I Learned

→Data privacy in grief tech is non-negotiable. Pet owners uploading photos of their late companions expect absolute security; we commit to purging files from active servers in 30 days and never using them for public training.
→Tone-sensitive AI engineering is key. Traditional LLM outputs can sound overly clinical or dramatic; prompt tuning was required to ensure the generated scripts remained humble, personal, and comforting.
→Orchestrating multi-model latency requires heavy job queuing. With image generation, text generation, voice-over compilation, and Hedra video generation all occurring sequentially, a robust asynchronous worker pool with real-time status updates was necessary to prevent timeouts.

// roadmap

Future Improvements

+A co-authoring feature allowing multiple family members to contribute memories
+Integration with local print-on-demand APIs to ship physical canvases directly
+An interactive memory wall for candles, virtual flowers, and family comments

// more.projects