Warehouse associates ship liquidation & donation inventory to vendors, school districts & charities.
I built a production-grade address book dashboard for exactly that.
#Nextjs 16 · #React 19 · #Figma MCP · 186 design tokens
Vibecoding CAN be production-ready. Here's how
Warehouse associates ship liquidation & donation inventory to vendors, school districts & charities.
I built a production-grade address book dashboard for exactly that.
#Nextjs 16 · #React 19 · #Figma MCP · 186 design tokens
Vibecoding CAN be production-ready. Here's how
Vibecoding with discipline = force multiplier.
AI handles translation.
You handle architecture, governance & craft.
The future of frontend isn't AI vs engineer.
It's AI × engineer.
👉 Live app: https://t.co/C47HbmH9oP
Who else is building production-grade with vibes? 👇
The result? Open source. Go read tokens.css.
Read the resolution log. Every decision documented.
Any new dev can open it and understand every design choice.
That's maintainability. That's scale.
👉 GitHub: https://t.co/0s1FPeuB7U
Step 5 — Figma MCP as a precision instrument.
→ Extract exact node IDs
→ Pull semantic token names from annotations
→ Cross-reference DS token docs
→ Flag where Figma deviated from the token spec
Design-to-code with receipts.
Step 3 — Systematic audit of hardcoded values.
Every magic number flagged.
The 7 remaining ones? Documented with WHY they couldn't map.
(content-driven widths, no DS token equivalent)
Principal engineers don't just fix — they LOG reasoning.
Step 2 — Token mapping as a first-class task.
430 lines. 186 tokens. 8 UI sections.
Every color, radius, spacing & typography value traced
back to a DS variable.
tokens.css = source of truth.
Not Tailwind hardcodes. Not inline styles.
Step 1 — Design system first, always.
Used @figma MCP to pull design context at the node level.
Not screenshots — actual component specs, spacing,
typography & radii piped directly into coding context.
👉 Figma: https://t.co/wkk5eSxtvI
The secret: you don't REPLACE engineering with vibes.
You ACCELERATE it.
AI handles repetitive translation.
You handle architecture, token hygiene & DS governance.
That's the real unlock. 🔓
The app: warehouse users flag inventory locations for
liquidation or donation.
They save multiple shipping addresses per location —
vendors, charities, school districts.
The dashboard decides WHERE it ships.
Real users. Real stakes. Real code.
@AdamMico1@RobertJanezic@_gringuinho_@salesforce@tableau@Benioff Stopped consulting and recommending Tableau to anyone for the past 2 years. The tools is heavily priced and the web UX is mediocre. The last couple version rollouts clearly shows data heavy updates over visuals. When was there was a major update on the standard visuals library?