Critiques

Honestly the second pass reads much calmer — the tighter measure alone fixed most of what bothered me in the first draft.

Choosing a bookshop as the subject was smart — it gives you a reason to show long-form text, and long-form is where a type system either holds or falls apart. The step from 15px body to a 44px display size is confident, and the shift to a warmer paper tone under the reading view is the detail I'd lead with in a portfolio walkthrough.

My one structural note is the "quiet mode" toggle. Framing comfortable line length and 4.5:1 contrast as a mode a reader has to switch on implies the default layout is allowed to be worse. WCAG-level contrast and a 66-character measure aren't a setting, they're the baseline — I'd rather see one reading view that already meets the target and drop the toggle entirely.

Second, the sample text does a lot of hiding for you. Every paragraph is roughly the same length, so the rag never gets tested. Drop in a real chapter with a two-word line, a long quoted passage and a footnote marker, then show what the hierarchy does when it's under pressure — that's the part hiring managers actually read for.

The bones are good and the topic is well chosen. Extend the flow with the checkout and receipt screens so the type scale has to survive dense numeric data as well as prose. Do that and this goes from a nice study to a case study.

Agreeing with the note above on the toggle, and adding one of my own: the caption style and the metadata style are 1px apart in size and identical in weight, so on the product page they read as one undifferentiated block. Pull the metadata down to 12px, drop it a step in colour, and let the caption keep the italic — you'll get the separation without adding a third family.

The baseline grid on the collection page slips by about 6px between the first and second column. Worth fixing before you share this around; it's the kind of thing a type-led review will catch in the first ten seconds.