Role: Founder, design, full-stack engineering
Stack: Next.js, Prisma + Postgres, NextAuth, Stripe, S3, fal.ai, Replicate
Models: FLUX LoRA training, Flux Kontext, Seedance, Kling and more
Year: 2025

OVERVIEW

Fine Photo Gen is a personalized AI photo and video studio. Upload 10 to 20 photos of yourself and it trains a custom FLUX model of you. Then generate unlimited, unique images of yourself in any scene, edit photos by describing the change, and turn stills into video.

No subscription required to start, results in minutes, and your training photos stay private.

FROM ONE IMAGE TO A SHORT FILM

STEP 01Generate a source imageOne picture, endless possibilities. Start from a prompt or from your own trained model.

STEP 02Create the storyboardScenes become panels and panels become your storyboard, powered by Flux Kontext.

STEP 03Animate & stitchEach panel becomes a video clip, and the clips are stitched together automatically.

Anime-style source image of an archer in a forest — The source image.

Storyboard panel 1 — Storyboard panels generated from that one image, ready to animate.

Storyboard panel 2 — Storyboard panels generated from that one image, ready to animate.

WHAT I SHIPPED

Selfie upload and personal model training (FLUX LoRA).
Photo editing with Flux Kontext (June 2025). Describe the change and it’s made.
Video generation (July 2025), with a picker covering current video models like Seedance Pro.
A credit system with Stripe subscriptions and one-off credit packs.
A redesigned, mobile-first landing page with a browsable demo library.

Demo (June 2025): picking a model, prompting, and generating an image.