GrowthLenzGembaFix
GembaFix by GrowthLenz

From Paper Logs to
Real-Time Intelligence

Digital process capture for PVC-O pipe manufacturing. Every parameter tracked, validated, and visible in real-time across all 6 production lines.

6

Production Lines

100+

Form Fields

14

Database Tables

Real-time

Data Sync

System Overview

Complete data flow from shop floor to executive dashboards

GembaFix System Overview - Manufacturing Process Digitization
System Architecture

Built for Scale

Modern stack designed for manufacturing environments

Shop Floor

6 TWINEX93 Production Lines

Line 1

Extrusion + Orientation

Line 2

Extrusion + Orientation

Line 3

Extrusion + Orientation

Line 4

Extrusion + Orientation

Line 5

Extrusion + Orientation

Line 6

Extrusion + Orientation

Digital Capture

Mobile/Tablet Entry with Validation

Extrusion Log

35+ fields

Orientation Log

25+ fields

Quality Check

22 checks

Downtime Log

16 codes

Inventory

6 materials

Changeover

15+ fields

Supabase Database

Real-Time PostgreSQL with Row-Level Security

chemfab_production_lines

chemfab_product_specs

chemfab_operators

chemfab_shifts

chemfab_extrusion_logs

chemfab_orientation_logs

chemfab_quality_inspections

chemfab_downtime_events

chemfab_downtime_reasons

chemfab_defect_master

chemfab_alerts

chemfab_bom_materials

chemfab_inventory_consumption

chemfab_changeover_logs

Role-Based Dashboards

Real-time Analytics by Permission Level

Admin

  • User & Role Management
  • System Settings
  • Plant Configuration
  • Audit Logs

Plant Head

  • Total Production vs Plan
  • Line-wise Comparison
  • Downtime Pareto
  • Yield Trends

Supervisor

  • Approval Requests
  • Shift Handover Summary
  • Escalations
  • Operator KPIs

Operator

  • Live Line Status
  • Enter Shift Readings
  • Report Downtime
  • Confirm Changeovers

Quality

  • Rejection % by Line
  • Defect Trends
  • Process Alerts
  • Rework Analysis
AI Intelligence Layer

Adding AI on Top of Your Data

The data we're capturing becomes the foundation for powerful AI. Two approaches - one available immediately, one that grows with your data.

LLM-Based Intelligence

Available now with existing data

Available Models (2026)

Closed-Source (API)

GPT-5.2Best reasoning, 400K context
Claude Opus 4.5#1 for coding, 200K context
Gemini 3 ProFastest, 2M context, multimodal
Grok 4.1Real-time search integration

Open-Source (Self-hosted)

GLM-4.7 ThinkingFrontier-level, MIT license
DeepSeek V3.290%+ quality at fraction cost
Llama 4 Scout10M token context
Qwen 3Strong multilingual support

Natural Language Queries

Ask questions in plain English

"What was Line 3's rejection rate last week?"

"Line 3 had 2.3% rejection last week, up from 1.8% prior. The spike occurred Thursday on Shift 2."

Auto-Generated Reports

Human-readable shift summaries

Generated Summary:

"Shift 1 produced 42 pipes with 97.6% acceptance. One 30-min downtime for material feed. Melt temp stable at 184°C."

Troubleshooting Assistant

Analyze data, suggest causes

"Wall thickness is varying. What should I check?"

"Haul-off speed fluctuated ±0.3 mpm in last 2 hours. Die temp zone 4 is 3°C below target."

Implementation Status

Coverage against the Process Digitization Proposal

Implemented in POC

Production Output Tracking
Downtime Reasons (16 codes)
Quality Parameters (Extrusion + Orientation)
Inventory Consumption (6 BOM materials)
Changeover Logging
Line-wise Data Entry (6 lines)
Shift-wise Validation
Total Production vs Plan
Line-wise Performance
Downtime Pareto Analysis
Live Line Status
Role-based Access (5 roles)
Process Deviation Alerts

Needs ChemFab Input

Other Product Sizes(Only DN400 provided)
KPI Definitions(Need from management)
Escalation Matrix(Need org hierarchy)
SCADA/PLC Capabilities(For machine integration)
Future Roadmap

Machine Integration Options

Three paths to evolve from manual entry to automated data capture

Phase 1Current

Manual Entry

Tablet/phone data entry validates workflows

  • Operator enters readings hourly
  • Form validation catches errors
  • Data flows to dashboard instantly
Phase 2

Hybrid Capture

OCR + selective machine integration

  • Scan existing paper forms
  • Connect machines with REST/MQTT
  • Operators verify auto-populated fields
Phase 3

Full Automation

SCADA/PLC direct integration

  • All readings from machines
  • Operators handle exceptions only
  • Real-time alerts on deviations

Machine Integration Architecture

SCADA/PLC

Modbus, OPC-UA, MQTT

Node-RED Gateway

On-site data transformer

Supabase

REST API ingestion

Dashboard

Real-time visibility

Node-RED runs on a Raspberry Pi or small server at the plant, bridging industrial protocols to cloud

Ready to Explore?

Start a production shift to enter data, or view the dashboard with sample data already loaded.