I built PetMemories.io to address a deeply personal challenge: the silence that follows the loss of a pet. While many of us have thousands of photos, they are often static. I wanted to see if generative video could help provide a sense of "final closure" by creating a peaceful, high-fidelity memory.

The inspiration came from observing a massive trend on TikTok. I noticed thousands of people were manually trying to create these "heavenly" tributes for their pets using basic filters, often with viral results. It was clear there was a strong B2C demand for high-quality memorial videos. I saw an opportunity to take this from a "manual social media hack" to a polished, AI-driven service that delivers professional results directly to a user's inbox.

The Tech Stack & Pipeline: I’m a big fan of keeping the "glue" simple so I can focus on the output quality.

Frontend/Orchestration: Built with Lovable for rapid iteration and Supabase for the backend.

The Vision Pipeline: I’m using Nano Banana Pro for the initial image synthesis because it handles animal anatomy and fur textures with much higher consistency than general-purpose models.

Video Generation: The final animation is handled by Veo 3. The main challenge was maintaining "character" (pet) consistency from the 2D photo to a 3D moving scene without it looking like a generic stock video.

I’ve tried to steer the prompts toward "peaceful/ethereal" rather than "hyper-realistic simulation" to avoid being emotionally jarring.

I’d love your feedback on:

The Business Model: Do you see other B2C verticals for this type of emotive generative video?

Output Quality: Does the video feel like a tribute or does it cross into the uncanny valley?

The Ethics: How do you feel about using generative media as a tool for grieving?

Happy to answer any questions about the integration with Veo or the workflow!