Kiln
A ceramics studio storefront rebuilt from a Dribbble concept, with every product photograph generated and graded rather than shot.
- Client
- Self-initiated — a ceramics studio concept, not a commissioned brief
- Role
- Solo — design, build, and AI photography direction
- Duration
- One week (Week 52), full-time evenings
How it was built
AI product photography — From a raw product shot to an editorial-grade image using a fixed prompt chain and post-processing pass.
The brief
The original Dribbble shot was a single static homepage mockup with no working store logic behind it. The brief was to take that concept to a fully responsive, purchasable storefront — and to solve the actual production problem a one-person ceramics studio has: no budget for a proper product photography shoot.
The build
- 01
Rebuilt the homepage layout at production fidelity first — grid, type, spacing — before touching any imagery, so the design decisions weren't compromised by placeholder photography.
- 02
Ran the AI product photography workflow (see the linked workflow below) against flat reference shots of the actual ceramic pieces, iterating the prompt chain until the lighting matched the brand's warm, gallery-lit aesthetic.
- 03
Wired the storefront's cart and checkout affordances against the template's existing component set rather than building bespoke commerce UI, keeping the template resellable as-is.
What didn’t work
The first two prompt iterations produced technically clean renders that read as obviously synthetic — too even, no material imperfection. The fix wasn't a better prompt; it was deliberately reintroducing asymmetry (uneven glaze pooling, slight throw-line waviness) as an explicit negative-prompt exclusion, which is documented in the linked journal entry.
Related Products
Related Prompts
Related Templates
Related Media
Related Journal Entries
Related Builds
- Kiln — a ceramics studio storefront
How it was made
- AI product photography