📋 Book Demo

/ Our Products

Innovative AI Vision Solutions
for Industrial Quality Control

Every product in our portfolio solves a specific manufacturing inspection problem — with zero false negatives, 80% lower cost, and deployment measured in hours, not months.

Zero False Negatives 99.8% Accuracy Any GigE Camera Edge Deployment

/ Full Product Portfolio

Built for Every Inspection Problem

Filter by need below — or scroll down for full specs on every product.

Core Vision Engine
CT Garud
CT Garud
AI-powered defect detection & segmentation — 10+ defect types in one frame, <50ms per inference.
Explore →
01
 No-Code Platform
CT Flow Builder
CT Flow Builder
Drag-and-drop AI inspection pipelines. No programming required. Multi-camera. Go live in hours.
Explore →
02
 OCR / OCV
CT OCR
CT OCR / OCV
Retrainable OCR for batch codes, expiry dates, CIJ, laser, and emboss prints at up to 400 PPM.
Explore →
03
 Pharma Specialist
CT Mikha
CT Mikha
100% blister pack and tablet inspection at 300 PPM. GMP audit trail. Pixel-level accuracy.
Explore →
04
 Unsupervised AI
Anomaly Detection
Trained on good parts only — flags any surface deviation. No defect labelling needed. Perfect for castings.
Explore →
05
 High Speed
BCFHS
High-speed batch code and print verification on FMCG lines at 1000+ PPM. Inkjet, laser, foil, emboss.
Explore →
06
 Metrology
Vision Measurement
Non-contact dimensional gauging at 20µm precision. Thread, slot, diameter, gap — all checked in one frame.
Explore →
07
 PLC Integration
Process Check (CPAS / NCMS)
Vision-guided process sequence interlock and manufacturing step verification with PLC/Modbus output.
Explore →
08
 Packaging
EOL Inspection
End-of-line packaging inspection for FMCG, pharma & beverage — seal, fill level, label, count verification.
Explore →
09
Not sure which product fits your problem?
Tell us what you're inspecting — we'll specify the right system and run a free proof of concept on your actual parts.
Get Free Assessment →
CT Garud
Core Product

CT Garud — AI Defect Detection & Segmentation

CT Garud is Crimson Tech's primary AI vision engine — a deep learning platform (YOLO-based) that detects, classifies, and segments 10+ defect types simultaneously in a single camera frame. Designed for real-time manufacturing inspection where zero defect escape is non-negotiable.

<50ms inference 10+ defect types / frame Any GigE camera Edge or Cloud PLC/Modbus output
Book CT Garud Demo →
✓ Frame #4312 processed in 23ms
✗ DEFECT: porosity 0.3mm² @ (412,218) → REJECT
✓ Frame #4313 — all 8 ROI clear → PASS
✓ PLC signal: OUT1=HIGH (PASS)
⟳ Frame #4314 queued · 3 cameras active
ℹ Shift report: 4,312 inspected · 9 rejected · 0 false neg
0
False Neg.
99.8%
Accuracy
23ms
Latency

Key Features

Multi-Defect Detection
Detects and classifies 10+ defect types simultaneously in one inference pass — porosity, cracks, scratches, missing features, dimension errors.
<50ms Per Frame
Real-time edge inference using TensorRT-optimized models. No cloud round-trip. Inspection decision before the part leaves the camera field.
Zero False Negatives
Trained and tuned for zero false-negative tolerance. Every defective part is caught. <0.2% false positive rate keeps production efficient.
Multi-Camera Ready
Orchestrate up to 8+ cameras in a single pipeline. Different camera positions, angles, resolutions — unified into one inspection decision.
PLC / Modbus Integration
Direct OK/NG signal to your existing PLC via Modbus TCP/RTU or digital I/O. No middleware. Auto-reject trigger on defect detection.
Retrainable by Your Team
Add new defect types, new product variants, or update acceptance criteria — in minutes. No vendor call. No extra cost. Your engineers do it.

Technical Specifications

Inference Speed<50ms per frame
Defect Types / Frame10+ simultaneously
Camera CompatibilityAny GigE Vision camera
PLC IntegrationModbus TCP/RTU · Digital I/O
DeploymentEdge (local GPU) or Cloud
OSWindows 10/11 · Ubuntu 20.04+
False Negative RateZero (guaranteed)
False Positive Rate<0.2%

Use Cases

  • Weld bead porosity and pin-hole detection on automotive components
  • Thread presence and slot dimension verification on machined parts
  • Surface scratch, dent, and contamination detection on metal stampings
  • Sticker, label, and badge position verification on assembled products
  • Black glue bead inspection on automotive glass panels
  • Bearing shell color band and slot verification
  • Multi-camera door seal and weatherstrip inspection
  • PCB solder joint and component orientation check

Frequently Asked Questions

CT Garud is an AI-powered defect detection and segmentation engine for manufacturing quality control. It uses deep learning (YOLO architecture) to detect surface defects, dimensional anomalies, presence/absence checks, and print quality issues in real time on production lines.
CT Garud supports any GigE Vision standard industrial camera — including brands like Basler, IDS, Hikvision, FLIR, and others. You are not locked into proprietary hardware. If you already have GigE cameras on your line, we can use them directly.
Yes. CT Garud processes a single camera frame and can classify 10+ different defect types in that frame — for example, detecting porosity, cracks, and missing features all in one inference pass, with separate OK/NG outputs for each.
A typical CT Garud deployment takes 1–5 days from hardware setup to go-live. This includes camera mounting, lighting setup, model training on your sample parts, and PLC integration. Retraining for a new product variant takes minutes.
Typically 200–500 images with labeled defect examples. Our team assists with image collection strategy and annotation. For anomaly detection mode, you only need good-part images — no defect labeling required.
CT Flow Builder
Key Differentiator

CT Flow Builder — No-Code Vision Pipeline

CT Flow Builder is a drag-and-drop AI inspection pipeline builder that lets your line engineers — not software developers — design and deploy complete multi-camera inspection systems in hours. Connect cameras, configure AI blocks, set PLC outputs, and go live — all without writing a single line of code.

Drag-and-drop builder Multi-camera support PLC/Modbus blocks Live dashboard Deploy in hours
See Flow Builder Live →
Drag-and-Drop Blocks
Camera input, AI detection, OCR, measurement, PLC output, reporting — all as draggable blocks. Connect them visually.
Multi-Camera Orchestration
Combine multiple camera views into one pipeline. Different AI models per camera. One unified inspection verdict.
Live in Hours
From zero to active inspection in hours. No software development. Your production team builds and owns the pipeline.
Retrainable AI Blocks
Swap or retrain AI models inside the pipeline without rebuilding the flow. New product = update the model block. Done in minutes.
PLC / Modbus Output Blocks
Trigger rejection gates, alarms, and conveyor controls directly from the pipeline. No custom integration code required.
Live Dashboard Included
Real-time OK/NG feed, shift statistics, defect trend charts, and last-NG image — auto-generated from your pipeline.

Frequently Asked Questions

CT Flow Builder is designed for production engineers, quality managers, and line supervisors who need to deploy AI vision inspection without software development skills. If you can use PowerPoint, you can build an inspection pipeline in Flow Builder.
Yes. Flow Builder supports multi-camera inspection pipelines — different cameras can run different AI models, and all results can be combined into a single OK/NG verdict for the part.
No programming is required. All inspection logic is built by connecting visual blocks. Our team sets up the initial pipeline and trains your team in one session — then your engineers own and maintain it independently.
CT OCR
OCR / OCV

CT OCR / OCV — Industrial Optical Character Recognition

CT OCR is a retrainable industrial OCR engine purpose-built for manufacturing batch code, expiry date, and print verification. Unlike general OCR, it handles metal-embossed text, low-contrast CIJ inkjet, laser engraving, and same-color printing at production speeds up to 400 PPM.

Up to 400 PPM CIJ, laser, metal, foil Any orientation Same-color detection Retrainable
Book CT OCR Demo →
✓ OCR: "MFG 2024-11-12" → PASS
✓ OCR: "EXP 08/2026" → PASS
✗ OCR: "BATCH A2X9" — missing char → REJECT
✓ OCR: "₹ 249.00 MRP" → PASS
⟳ 400 PPM · 99.8% verification rate · 0 false release
400
PPM Speed
5 types
Print Technologies
5 Print Type Support
CIJ inkjet · laser engraving · metal emboss · foil print · thermal transfer — all in one retrainable engine.
Any Orientation
Reads rotated, skewed, and multi-line text without pre-alignment. Works on curved surfaces like bottles and cans.
Same-Color Detection
Detects same-color-on-same-color embossed text that standard OCR systems fail on completely.
Expiry Date Validation
Logic-level date validation — catches transposed digits, future/past date errors, and format violations automatically.

Use Cases

  • Batch code and expiry date verification on FMCG packaging (pouches, bottles, cartons)
  • MRP price verification on retail packaging
  • Laser-engraved serial number reading on metal components
  • CIJ inkjet date code inspection on beverage cans and bottles
  • Embossed batch code verification on pharma blister packs
  • Multi-line label text verification at up to 400 PPM
CT Mikha
Pharma Specialist

CT Mikha — Pharmaceutical Inspection System

CT Mikha is Crimson Tech's dedicated pharmaceutical vision inspection system. It delivers 100% blister pack and tablet inspection at up to 300 PPM with pixel-level accuracy, GMP-compliant audit trails, and automatic rejection of ALU-ALU, foil, and PVC blister defects.

300 PPM ALU-ALU & foil GMP audit trail Pixel-level accuracy 100% inspection
Book CT Mikha Demo →
CT Mikha — Blister Inspection
300
PPM Inspection Speed
100%
Blister Coverage
✓Missing tablet detection
✓Broken / cracked tablet
✓Wrong color / shape
✓Foil seal integrity
✓GMP audit trail
100% Blister Inspection
Every blister pocket inspected — no sampling. Missing, broken, discolored tablets all caught before packaging exit.
GMP Audit Trail
Every inspection result logged with timestamp, batch, image capture, and OK/NG verdict. GMP-compliant records for regulatory submissions.
ALU-ALU & Foil Support
Works on both standard PVC/foil blisters and harder-to-inspect ALU-ALU opaque packs using specialized imaging techniques.
Pixel-Level Accuracy
Sub-millimeter detection of cracks, chips, and surface defects on tablets. Color and shape classification per tablet.

Frequently Asked Questions — CT Mikha

CT Mikha inspects PVC/foil blister packs, ALU-ALU blister packs, tablet strips, and capsule blisters. It checks each pocket for missing, broken, cracked, discolored, or wrong-orientation tablets, plus foil seal integrity.
CT Mikha detects: missing tablets, broken tablets, cracked tablets, chipped edges, wrong color, wrong shape, double-fill (two tablets in one pocket), wrong orientation, and foreign particle contamination in the blister pocket.
Yes. CT Mikha generates a complete inspection log per batch — including timestamps, product details, inspection images, OK/NG results, and operator information — compliant with GMP documentation requirements.
Unsupervised AI

Anomaly Detection System — Learn From Good Parts

The Crimson Tech Anomaly Detection System uses unsupervised learning — trained exclusively on good-part images, with no defect labeling required. It builds a model of what "normal" looks like, then flags any surface deviation that falls outside that normal range. Perfect for complex castings, stampings, and machined parts where defects are unpredictable.

No defect labels needed Train on good parts only Surface + texture Any material
Book Anomaly Detection Demo →
// Training mode: 250 good-part images loaded
✓ Model: normal surface pattern learned
// Inference mode active
✓ Part #1201 — deviation score: 0.02 → PASS
✗ Part #1202 — deviation score: 0.87 @ (320,180) → ANOMALY DETECTED
⟳ Anomaly map generated · PLC reject signal sent
No Defect Labeling
Train on 200–500 good-part images. The AI learns what normal looks like — no defect examples needed, no annotation required.
Finds Unknown Defects
Catches defect types you didn't pre-define — porosity, inclusions, unexpected surface texture changes, contamination.
Anomaly Heat Map
Generates a spatial heat map showing exactly where on the part surface the anomaly was detected — for operator review and root cause analysis.
Variable Part Geometries
Ideal for die castings, investment castings, and forged parts where geometry varies slightly part-to-part. The model adapts to natural variation.
High Speed · FMCG

BCFHS — Batch Code & Print Verification at 1000+ PPM

BCFHS (Batch Code and Fill Height Sensor) is Crimson Tech's high-speed print and batch code verification system, designed for FMCG packaging lines running at 1000+ parts per minute. Verifies batch codes, expiry dates, barcodes, and fill heights at production line speed — zero false releases.

1000+ PPM Inkjet · laser · foil · emboss Fill height sensing Zero false releases
Book BCFHS Demo →
1000+
Parts per minute verified
Inkjet CIJ✓ Supported
Laser Engraving✓ Supported
Foil Print✓ Supported
Embossed Text✓ Supported
Metrology

Vision Measurement System — 20µm Dimensional Gauging

Crimson Tech's Vision Measurement System replaces contact gauges and CMMs with non-contact AI vision measurement. Achieves 20µm measurement precision for thread pitch, slot depth, diameter, gap, and surface profile — all in a single camera frame, at production line speed.

20µm precision Non-contact Production speed Thread · slot · diameter
Book Measurement Demo →
Measurement Precision20µm (0.02mm)
MethodNon-contact optical
Measured FeaturesThread, slot, diameter, gap, profile
OutputDimensional value + Pass/Fail vs tolerance
IntegrationSPC data export · Modbus · CSV
Throughput100% inline at production speed
PLC Integration

CPAS / NCMS — Vision-Guided Process Verification

The Crimson Poka-Yoke Assistance System (CPAS) and Neutral Channel Management System (NCMS) use machine vision to verify manufacturing process steps — ensuring the right operation was performed, the right component was installed, and the process sequence was followed — before the line is allowed to advance.

Process interlock Step verification PLC hold signal Poka-yoke
Learn More →
  • Verify correct sub-assembly was installed before next operation is permitted
  • Check process sequence compliance — prevents assembly in wrong order
  • Confirm torque markings, paint marks, and process signatures are present
  • Vision-guided robot confirmation before arm advances to next position
  • Process step photo documentation for traceability
Packaging

EOL Inspection — End-of-Line Packaging Quality Control

The EOL (End-of-Line) Inspection System is a comprehensive packaging line inspection solution for FMCG, pharma, and beverage manufacturers. It checks every unit leaving the line — seal integrity, fill level, label presence and position, cap and closure, barcode, and count — before shipping.

Seal inspection Fill level Label & cap check Count verification FMCG · Pharma · Beverage
Book EOL Demo →
  • Seal integrity check — heat seal failures, open packs, weak seals
  • Fill level detection — underfill and overfill by volume
  • Label presence, position, and orientation verification
  • Cap and closure torque mark and presence inspection
  • Barcode and QR code readability verification
  • Case count and pack configuration verification at shipper level
  • FMCG sachet seal and noodle sachet completeness check
  • Beverage can and bottle fill level + crown seal inspection

/ Get Started

Not Sure Which Product
Fits Your Line?

Tell us what you're inspecting. We'll run a free proof of concept on your actual parts — no commitment. Within 1 week.

Book Free POC Assessment → See Industry Use Cases