Carry Go — Courier App

Overview#

The Courier App is the field tool for Carry Go's delivery agents: a Flutter app covering shift attendance, pickup and delivery requests, inter-warehouse transfers and fleet-asset tracking, all backed by the same Frappe/ERPNext data model as the Warehouse Portal.

Problem#

Delivery agents had no reliable way to prove attendance for payroll purposes, no structured workflow for accepting and completing pickups, and no way to hand off packages between warehouses without a phone call and a manual note. Fleet assets like scooters had no tracking beyond memory.

Solution#

I designed an HMAC-signed QR attendance system — an agent scans a per-shift code that's cryptographically verified server-side, closing the gap that a simple photo or GPS check-in would leave open. Pickup, delivery and inter-warehouse transfer requests all flow through explicit, state-tracked screens rather than free-text notes, and every fleet asset gets a scannable QR code tied to its ERPNext record.

Architecture#

The app talks to the same Python/Frappe REST layer as the Warehouse Portal and Customer App, running on ERPNext v16's isolated Bench. Firebase Cloud Messaging pushes shift and order events to the agent in real time; HMAC signing on the attendance QR flow prevents replay or spoofing of clock-in events.

Features#

• ERPNext-backed authentication and profile onboarding for new agents.
• HMAC-signed QR attendance for shift clock-in/out.
• Pickup-request acceptance with customer and locker details.
• Inter-warehouse transfer missions with handoff confirmation between couriers.
• Fleet asset tracking (e.g. scooters) via scannable QR codes.
• FCM push notifications for shift and order events.

Development Process#

The app was built directly against the Frappe REST endpoints I designed for the Warehouse Portal, so agent-facing actions (accepting a pickup, completing a transfer) write to the exact same records supervisors see live in the web portal — no separate sync layer, no drift between what the agent sees and what the office sees.

Challenges#

Getting attendance verification right without adding friction for agents in the field was the core design problem — GPS alone is spoofable and a manual photo check-in is slow. The HMAC-QR approach keeps clock-in to a single scan while remaining cryptographically verifiable server-side.

Results#

Delivery agents now run their entire shift — clock-in, pickups, transfers, asset handling — from one app, with every action immediately visible to warehouse supervisors through the shared backend.

Future Improvements#

Offline support for low-connectivity warehouse zones, and in-app route guidance between pickup and delivery points.