Perception at line speed
Metropolis and DeepStream with TensorRT and Holoscan drive high-speed print-defect, color, register, structural and crush, warp and glue and fold vision.
Technology
A plant-edge autonomy stack, grounded and auditable
Perception models, process agents, bounded control policies, a physics-informed twin and a closed-loop data flywheel, built for the physics and brands that converting lives and dies on.
Seven agents under one plant orchestrator
Fine-tuned models
Fine-tuned defect, register and color, structural and warp and fold-quality vision plus corrugator and waste and OEE optimization models at the edge; time-series prognostics for machines.
Reasoning where it pays
A model router serves the best or cheapest model per step; high-volume steps use fine-tuned open models, premium root-cause reasoning uses frontier models.
Bounded control
Process agents issue setpoint moves within approved limits, with human-in-the-loop checkpoints and graduated autonomy.
NVIDIA physical-AI stack, edge to twin
Foldeon runs a closed perceive-decide-act loop across the converting plant, from Jetson at the edge to Omniverse for the twin.
- Isaac and Jetson for robotic feeding, stacking, palletizing and changeover
- Metropolis and DeepStream with TensorRT and Holoscan for high-speed vision
- Triton and NIM serving; NeMo-fine-tuned reasoning models
- cuOpt for scheduling, nesting, makeready and palletizing
Edge cell, Jetsonloop latency
Defect decision p95< 50 ms
Register advisory p95< 20 ms
Camera streams per line4 to 12
ServingTriton and NIM
A converting-line-and-box twin that runs the job before you do
An Omniverse twin simulates corrugation, print and die and fold to hit target quality and yield before the run, auto-optimizing board, print and fold and validating every change first.
- Physics-informed twin of corrugation, print, die and fold and handling
- Pre-run job simulation and makeready auto-tuning
- Cosmos and Omniverse Replicator synthesize rare defect scenarios
- Isaac Sim and Lab validate robot cells before deployment
The closed AI workflow
Perceive, plan, run, sense and predict, optimize, flag, log, retrain.
Perceive
Sense board, sheet and box across every stage with fused vision and sensors.
Plan
Plan corrugate and bond, print and color, die-cut and crease and fold and glue for the job.
Run adaptively
Execute with adaptive, bounded closed-loop control.
Sense and predict
Predict defects, color, structural strength, warp and fold quality.
Optimize and flag
Optimize waste, makeready and OEE; flag defect and claim risk.
Log and retrain
Log immutably, then retrain the fleet on production and synthetic data.
What powers each layer
| Capability | Technology |
|---|---|
| Edge vision and fusion | Jetson Orin and Thor, TensorRT, Holoscan, Metropolis and DeepStream |
| Serving | Triton and NIM at plant edge and private cloud |
| Reasoning and RAG | NeMo with Retriever and Guardrails, pgvector, citations enforced |
| Training | DGX and HGX, 10k to 40k GPU-hours per major release |
| Simulation | Omniverse and OVX twin, Cosmos and Replicator synthetic data, Isaac Sim and Lab |
| Optimization | cuOpt: sequencing, nesting, makeready order, energy, pallet patterns |
Trust, safety and evaluation, built in
Grounded outputs are non-negotiable on a plant floor.
Grounded and cited
Citation and grounding checks on every output; RAG enforces traceable sources.
Human-in-the-loop
Graduated autonomy with approval checkpoints for high-risk changes.
Continuously evaluated
Golden datasets and LLM-as-judge gate every model and prompt change in CI.
Isolated and auditable
Per-tenant isolation, strict IP, artwork and recipe protection, immutable audit log.
Defect decision p95
< 50ms
Register advisory p95
< 20ms
Synthetic scenarios
1Mper quarter
Camera streams
4 to 12per line
Per-plant history and per-craftsman memory
Foldeon captures the scarce craft of press operators, corrugator crews and die-makers as versioned, tenant-scoped memory that improves over time without cross-tenant leakage.
- Per-plant converting-quality-and-waste history
- Per-operator and die-maker performance memory
- pgvector defect, color-image and job and recipe retrieval
- Time-series store for corrugator, press and gluer telemetry
The router and the twin are the two ideas that make this economical: cheap models where volume is high, and a simulation that pays for makeready before the run.
Questions converters ask
Can we run fully on-prem?
Yes. Optional on-prem deployment is available for sensitive converters, with per-tenant isolation and strict IP, artwork and recipe protection throughout.
How do you control inference cost?
Models are abstracted behind a router so high-volume steps move to fine-tuned open models to control COGS, while premium reasoning steps use frontier models only where they add value.
How do you handle rare defects with little data?
Cosmos and Omniverse Replicator synthesize 100k to 1M rare defect, material, color, fold and handling scenarios per quarter to augment production data.
Go deep with our engineers
Bring your line topology and machine list. We’ll map the edge, twin and integration plan.