# Klyff Product Suite

## metadata
- title: Manufacturing Intelligence Products | Klyff
- route: /product
- url: https://klyff.com/product
- format: product_overview

## Platform Overview
Klyff combines four interoperable product pillars for manufacturers that need a practical path from disconnected plant signals to dependable operational decisions. Each module can start independently on a single line, then connect to the others as requirements grow. The shared edge foundation helps teams reuse device integrations, operating context, and deployment practices instead of rebuilding a separate stack for every use case.

The suite supports operations leaders who need a live plant view, quality teams scaling inspection, reliability teams prioritizing maintenance, and enterprise teams improving performance across plants without centralizing sensitive raw data.

## Product Pillars
The four modules work together as one manufacturing intelligence platform, while each can also address a focused plant need on its own. Start with the data layer, inspection, reliability, or fleet learning that matches the operational goal, then connect other capabilities using shared integrations and context.

### Analyzr — Unified Plant Telemetry and Control Plane
Analyzr bridges heterogeneous shop-floor hardware into a standardized data fabric. It connects legacy PLCs and modern sensors through industrial protocols including OPC-UA, Modbus TCP/RTU, MQTT, EtherNet/IP, and Profinet, then streams normalized time-series data to local analytics engines.

#### Capabilities
- **Multi-protocol translation:** Connect existing plant equipment without a rip-and-replace project.
- **Edge buffering and data continuity:** Buffer data locally through network interruptions to reduce data loss.
- **Real-time anomaly streaming:** Surface sub-second signals, thresholds, and statistical process control (SPC) views.
- **Clean API exports:** Provide normalized metrics to local SCADA, MES, and downstream AI services.

Use Analyzr to unify machine and sensor telemetry, monitor OEE, first-pass yield, alarms, throughput, energy, and asset health, and supply trusted operational signals to inspection and maintenance applications.

### Inspeqtr — Visual Quality Inspection at Line Speed
Inspeqtr runs optimized deep-learning vision models on industrial edge hardware such as NVIDIA Jetson, Intel Xeon, and industrial PCs. It detects surface blemishes, dimensional deviations, weld porosity, and missing components in real time using production camera feeds.

#### Capabilities
- **Low-latency inference:** Generate defect bounding boxes and segmentation masks for high-speed production lines; the product page describes sub-10ms inference capability.
- **Synthetic defect augmentation:** Help train useful models when examples of rare defects are limited.
- **Closed-loop line integration:** Send digital I/O signals to reject mechanisms or line controls when the validated workflow calls for it.
- **Active learning feedback:** Let operators label edge cases so local models can be reviewed and improved.

Use Inspeqtr to increase inspection coverage, identify defects consistently, preserve evidence for review, and integrate inspection outcomes into operator workflows.

### Prescptr — Prescriptive Maintenance Intelligence
Prescptr turns high-frequency vibration, thermal, and electrical telemetry into remaining useful life (RUL) projections and maintenance guidance for critical equipment.

#### Capabilities
- **Vibration and spectral analysis:** Apply continuous FFT and wavelet analysis to signals from bearings, pumps, gearboxes, and similar assets.
- **Prescriptive guidance:** Provide recommended repair protocols and spare-parts context alongside anomaly alerts.
- **Context-aware baselines:** Account for operating speeds and recipe changes to help reduce false positives.

Use Prescptr to surface changing asset condition, prioritize maintenance by risk, and support planned interventions before equipment issues interrupt production.

### Senatr — Multi-Site Federated Learning
Senatr coordinates distributed edge learning across manufacturing sites. Each plant trains models locally on its proprietary sensor feeds, while encrypted model updates—not raw images or telemetry—are shared with the global aggregation process.

#### Capabilities
- **Local data control:** Keep raw production images and telemetry within plant premises.
- **Fleet-wide model improvement:** Combine learnings across multiple sites so validated improvements can benefit the wider fleet.
- **Heterogeneous hardware support:** Support mixed edge environments such as Jetson, x86, and TPU deployments.

Use Senatr when plants need to improve models across different equipment and operating conditions while respecting data residency, trade-secret, and compliance requirements.

## Shared Deployment Approach
Klyff connects to existing OT equipment, validates models against production conditions, deploys inference on suitable edge hardware, and gives operators a workflow for reviewing exceptions. Modules can be adopted separately or combined over time; shared telemetry and fleet practices make later expansion easier.