A Porter-style logistics platform: customer, driver and fleet-manager apps plus a super-admin dashboard, built AI-first.
Overview
Shipton is a full logistics product in one monorepo: a documented backend and four frontends. It's my test of how far an AI-driven workflow goes. Designs live in Figma, Claude Code writes the code, and I direct, review and own every decision.
My role
Sole frontend developer- 01Sole frontend developer across four apps: customer, driver, fleet manager and super admin.
- 02Designed the monorepo layout: backend/ plus frontend/ with one folder per app.
- 03Run Claude Code in VS Code against the codebase and backend docs.
- 04Match the Figma file through Figma MCP, component by component.
How it was built
AI-drivenFully AI-driven. I write the intent, constraints and reviews; Claude Code writes the code against the backend docs and the Figma file.
- Figma
Design
Every screen for all four apps lives in the Shipton Figma file.
- Figma MCP
Extract
Frames, tokens and specs pulled into the editor, no Dev seat needed.
- Claude Code
Generate
Screens and components written against the documented backend APIs.
- Me
Review
Architecture calls, code review, type checks and fixes.
- Expo Go
Test
Run on a real phone over local IP to check every flow.
Key features
Customer app
Book a delivery, track it, pay.
Driver app
Accept jobs, navigate, update status.
Fleet manager app
Manage drivers and vehicles as an agent.
Super-admin dashboard
Web control panel for the whole operation.
Impact
Screens
12 of 13 addedStack
Mobile
Web
AI workflow
Backend
talk.