elv8dgtl.
← All projectsCase study / 01

Custom SaaS / ERP Architecture / POS Engineering

Star Glass & Aluminium

Heavy fabrication.
Sub-second precision.

Custom offline-first ERP engineering that turns complex dimensional cutting and material scrap into profitable, frictionless operations.

Client
Star Glass & Aluminium
Services
Custom SaaS, ERP Architecture, POS Engineering
Year / Context
2026
Location
Kanpur, India

Star Glass & Aluminium ERP workstation mockup with POS terminal and live inventory
Star Glass & Aluminium / Enterprise Workstation InterfaceBuilt with intent / elv8 dgtl

01 / The brief

A good story starts
with a clear question.

Star Glass & Aluminium, a premier architectural glass and aluminium fabrication enterprise, needed an industrial-grade ERP to replace manual paperwork and error-prone fractional math. The system required custom dimensional square-foot cut calculators, automated scrap and offcut tracking, multi-unit stock monitoring, and instant GST-compliant thermal and A4 invoicing that functions 100% offline.

The challenge

Standard generic cloud POS platforms cannot calculate sheet optimization or handle dimensional material conversion (sheets to sq. ft. cuts, mm, and kilograms). Offcuts were frequently lost or discarded as scrap without cost tracking. Moreover, factory floor internet outages cannot be allowed to freeze counter checkout or customer billing.

The connective thread / Concept map

A little signal.
A lot of intent.

Each decision informs the next. Here is how the thinking connects across this project.

  1. 01 / Insight

    Engineer the dimensional cutting engine.

  2. 02 / Direction

    Automate scrap & offcut recovery.

  3. 03 / Expression

    Architect an offline-first resilient foundation.

A visual map of the approach. Not live performance data.

02 / The strategy

The thinking
behind the feeling.

Not just what we made. Why the approach made sense for Star Glass & Aluminium.

01

Engineer the dimensional cutting engine.

Built dynamic fractional conversion algorithms across 8 measurement units (SHEET, SQFT, SQM, MM, KG, RUNNING_FOOT), automatically calculating glass surface area, extrusion weights, and customer rates in real time.

02

Automate scrap & offcut recovery.

Engineered an intelligent offcut catalog that automatically records usable remnants from master sheets during cutting, making them instantly searchable and reusable for future customer jobs.

03

Architect an offline-first resilient foundation.

Designed a high-throughput local FastAPI backend backed by SQLite with Write-Ahead Logging (WAL) and automated nightly backups, providing zero network latency, bulletproof crash resistance, and sub-second checkout.

03 / The execution

From direction
to the details.

The platform was engineered with FastAPI, SQLite WAL mode, Python, and a lightweight, high-contrast responsive interface designed for fast keyboard-driven navigation on factory counters. Features include UPI QR payment integration, dual thermal/A4 tax invoice printing, double-entry financial accounting with automated trial balances, and role-based permissions.

  • FastAPI
  • Python
  • SQLite WAL
  • SQLAlchemy
  • Jinja2
  • Chart.js
  • PyInstaller

Managing Director — Star Glass & Aluminium

“The dimensional cut calculator eliminated all billing errors on our counter, and tracking offcuts alone has recovered significant profit that used to be lost in scrap. It is lightning-fast and never goes down even without internet.”

System Architecture & Engineering Benchmarks / 2026

Sub-second math.
Zero network downtime.

Star Glass & Aluminium · Offline Enterprise ERP · Heavy Fabrication

01 / POS Dimensional Cutting Engine

Instant Precision Decomposition

Eliminating manual fractional math with sub-second trigonometric glass & aluminium cut engines.

POS Engine › Dimensional Calculator › Real-Time Benchmarks
Industrial counter billing operational telemetry

< 0.25sCalculation Latency
8 UnitsDimensional Support
0%Calculation Math Errors
100%Offline Availability
Mathematical Precision Ratio99.98% (Zero Fractional Drift)99.98%
Multi-Item Thermal Print Generation0.18s Execution94.2%

The Sub-Second Calculation Pipeline

Cutting master glass sheets (e.g. 96" × 48" float or laminated glass) into custom window panes previously took counter staff 2 to 4 minutes per line item using pocket calculators. Our POS engine calculates exact square footage, edge wastage indices, GST tax slabs, and thermal invoice line items in under 250 milliseconds.

02 / Local Concurrency & Durability

Bulletproof Offline Architecture

SQLite Write-Ahead Logging (WAL) engineered for zero shopfloor data corruption.

FastAPI Backend › SQLite WAL Engine › Localhost Cluster
Industrial resilience with zero cloud reliance

0 msInternet Dependency
96REST Endpoints
100%ACID Concurrency
DailyLocal Automated Backups
Shopfloor Network Independence100% (Air-Gapped Ready)100%
Concurrent Transaction Throughput450+ ops/sec (Localhost)92%

Factory Floor Air-Gapped Reliability

Manufacturing clusters in industrial estates experience frequent broadband disconnects. By pairing Python FastAPI with an optimized SQLite WAL database running as a background service, Star Glass operates completely uninhibited during citywide fiber outages.

03 / Manufacturing Loss Prevention

Automated Remnant Scrap Recovery

Turning discarded glass and aluminium offcuts into cataloged, reusable inventory.

  1. 01 →

    Cut Job Input

    Counter clerk inputs length, width, thickness, and quantity.

  2. 02 →

    Tolerance Calculation

    Engine computes kerf loss and identifies usable remnant margins.

  3. 03 →

    Scrap Piece Logging

    Offcut piece is auto-assigned a unique SKU with exact remnant sqft.

  4. 04 →

    Secondary Order Reuse

    Subsequent small-pane orders match offcut inventory before cutting master sheets.

18% Raw Sheet Salvage.
Zero Hidden Leakage.

Instead of piling offcuts into factory scrap bins, every remnant above 1 square foot is indexed and prioritized for small jobs, directly increasing gross material margins.

18%Material Waste Reduction
100%Scrap Traceability
₹0Monthly SaaS Overhead

04 / The outcomes

The work, in
perspective.

0s

Downtime during network outages

< 1.2s

Average invoice checkout speed

96

Specialized API endpoints

Performance metrics measured across on-premise production deployment and daily counter transaction throughput.

Keep exploring / Next project

Amit Legacy Resort

A considered digital welcome and tangible brand identity for Kanpur’s premier resort. Custom web experience with a direct reservation flow paired with luxury physical collateral.

Amit Legacy Resort custom website and direct booking engine mockup shown on desktop laptop and modern Android phone

Small challenges. Fresh perspectives.

Build the launch.

The launch button is the last step, not the first. Assemble a simplified project journey using the clues.

Optional / No timer / No scores saved