Service 04 / Web Development
Loading isn't shipping.
Searchable and sellable is.
A site that runs in production, not a design file.
Build, payments, SEO foundations and deployment in one pass,
handed over so you can edit copy and add pages yourself.
Who this is for
- ·You have a product or service to sell and need site + payment + delivery
- ·Your current site is templated and every copy change goes through the vendor
- ·You need a bilingual site and worry two content trees will be unmanageable
- ·The site exists but performs badly, breaks on mobile, or doesn't get indexed
Does this sound familiar?
「We need a site but don't know which stack; we're afraid of being stuck after launch」
「We can't edit the template ourselves; every text change costs money」
「Our English pages show up in search with Chinese titles」
「Lighthouse scores are poor and it takes seconds to open on a phone」
What you get
What I don't do
Stating the limits saves both sides more time than listing capabilities.
Not a design studio
If you have a designer, I implement the design. I don't do brand identity or logos — I'll refer you on.
No low-code platforms
No Wix, Squarespace or Webflow. They lock you in, and structural changes follow their rules — higher long-term cost.
No full e-commerce builds
No inventory, multi-SKU, or logistics. I do 'a few digital products plus Stripe' and that's the edge of it.
No aesthetic guarantees
Structure, hierarchy and performance are mine to guarantee. Visual taste is subjective; we align on a real sample page during scoping.
Timeline and pricing
Fixed pricing, not hourly. Under $2,200: 50% upfront, 50% on acceptance. Above: three instalments. Third-party costs (server, API, domain) are not included.
Why take my word
This site is the output of that process
Next.js 16 + React 19 + Tailwind, 34 URLs, bilingual. Lighthouse: 97 performance, 100 accessibility, 100 best practices, 100 SEO.
Bilingual language tags are genuinely hard
In App Router, html lang can only be defined once. Switching per route requires splitting route groups. Four approaches and three wrong ones are written up.
Shipped a 126-page site before
The previous version had 126 pages on a mixed SSG/ISR architecture with 14 Supabase tables. Scale isn't the problem; unclear structure is.
Performance claims are measured
Images went from 955 KB to 105 KB. Removing a blocking script took TBT from 240ms to 60ms. Recorded, not estimated.
Common questions
Which stack, and why not WordPress?
Next.js with TypeScript and Tailwind. Chosen for controllable performance, SEO friendliness and low hosting cost. WordPress suits very large content operations needing visual editing, at higher performance and maintenance cost.
Can I edit it myself afterwards?
Yes. Content is Markdown — editing copy is editing text. Components are structured so adding a page means copying one and changing its data. The handover recording walks through it.
How long to launch?
15–20 working days for Standard, 20–25 for Full. That includes your review time; if assets are late, the timeline shifts.
Do I need to buy a server?
Yes, but it's cheap: a Hong Kong VPS runs the site, scheduled jobs and database for tens of RMB a month. Third-party costs aren't in the quote.
Can we build one page first to check?
Yes. During scoping I produce one real page — not a mockup — so you can confirm style and structure before the rest is built.
Tell me what you're selling
What it is, who it's for, roughly how many pages, and whether it needs to be bilingual. That's enough for a quote and timeline — no requirements doc.
Describe the projectOther services
Public pricing, written boundaries, and an honest no when it's not a fit. That saves more time than any promise.