Connectors
Spinning, weaving and knitting, dyeing and finishing, inspection and colour-MES connectors for the wedge workflow. Read paths first; write paths behind the envelope.
Changelog
Every release that can touch a machine, written down. Starting before the first one.
Nothing has shipped to a customer mill yet. This page holds the pre-release build log, the roadmap from our own plan, and the policy a release will follow before it is allowed near a setpoint. Entries carry a status, not a date, until something ships.
Build log
What is being built for the first release
The MVP scope from our plan, component by component. "In build" means code exists against the wedge workflow. "In design" means the architecture is settled and the work has not started.
Agent loop
A constrained, production-grade sense, reason, act loop with human-in-the-loop checkpoints on every control action.
Review console
The console where textile engineers and colourists approve, correct and give feedback. Corrections become training signal.
Audit log
Assurance-grade, append-only quality and sustainability record from day one, exportable for buyer and regulatory audit.
Mill edge runtime
Jetson-class inference at the cell, a GPU server per mill for multi-model serving on Triton, offline fallback.
Vision pipeline
DeepStream and TensorRT pipelines for slubs, holes, streaks, stains, floats, barré, skew, shade drift and width faults. Sub-100 ms per frame is the design target. ASPIRATIONAL
Fabric twin
An Omniverse-based as-produced twin that qualifies recipes, tensions and sequences before a machine moves. ASPIRATIONAL
Policy engine
Autonomy level, approval conditions, setpoint envelopes and revert conditions declared alongside each agent as code.
SDKs and API
Python and TypeScript SDKs, streaming run events, webhooks and a mill simulator for development against a virtual line.
Policy
How a release reaches a machine
Once there is a release, this is the path it takes. Write-path changes never skip a step.
-
Phase 1
Publish
Release notes name every write-path change and attach the replay diff against the previous version.
-
Phase 2
Replay
The mill replays its own recent runs against the candidate and reviews the decision diff.
-
Phase 3
Shadow
The candidate runs in shadow on one machine while the pinned version stays in control.
-
Phase 4
Promote
For write-path changes the site signs off explicitly. The pin moves at a step boundary, never mid-step, and one command rolls it back.
Roadmap
What comes next, from our own plan
Every milestone below is aspirational until contracts, deployed connectors or measured mill results exist. We would rather label that than imply it.
| Horizon | Milestones |
|---|---|
| 0 to 6 months | Three to five design partners on the wedge workflow. MVP shipped: connectors, agent loop, review console, audit log. Accuracy and seconds, re-dye and waste ROI proven in shadow, then assist mode. SOC 2 Type I in progress. ASPIRATIONAL |
| 6 to 12 months | Design partners convert to paid. Graduated autonomy for low-risk inspection, dye and finishing loops. Twin hits shade before the run. A second module in the same mill. SOC 2 Type II, SSO and RBAC. ASPIRATIONAL |
| 1 to 3 years | The full spin, weave or knit, dye, finish and inspect loop end to end. Multi-site textile groups. Machinery and dyeing-systems vendor channel. Benchmark data products as premium tiers. ASPIRATIONAL |
| 3 to 10 years | The autonomous system of action for textile manufacturing globally, with a platform and API ecosystem across adjacent scopes through to the cut-and-sew handoff. ASPIRATIONAL |
Versioning
Pins, support and rollback
Each mill pins its own version
A mill runs the version it has signed off on, not the latest. Updates are proposed, replayed and shadowed before the pin moves, and the previous pin stays installed for rollback.
- Per-mill version pin, moved only by the site after sign-off
- Advisory-only changes may be promoted by the platform owner; write-path changes never
- Previous version kept installed; rollback is one command and is logged
- API versions date-pinned per key, so breaking changes never arrive on their own
- Model artefacts signed and verified on load
Guiding principle
Trust before autonomy
Grounding, audit, human-in-the-loop and graduated control, in that order. Every supervised correction trains the system.
Land
Narrow
One wedge workflow.
Expand
Relentlessly
Lines, modules, sites, autonomy level.
Run timeline
The reference scenario each release is replayed against
Run RUN-4471 is the scenario this site uses everywhere. Each release will be replayed against it, and against each mill’s own history, before the pin moves.
RUN-4471
RUN COMPLETE
Scenario mill · reactive exhaust jet dyeing · illustrative run, not a customer deployment
Goal: Dye lot DL-4471 · 3,180 kg · 40s combed cotton poplin · shade Indigo 19-4028 TCX · ΔE ≤ 0.8 · due Thu 06:00
-
01 ingest.lot_state SUCCEEDED 0.8s
Pulls greige genealogy for 14 rolls: yarn lots, loom, greige GSM 138, absorbency and residual size from the prep line.
Agent, input and output
agent Quality-and-Conformance result ← 14 rolls · 3,180 kg · mean GSM 138.2 (σ 1.1) · prep OK -
02 twin.simulate_lot SUCCEEDED 46s
Omniverse fabric-line twin simulates the exhaust curve, machine loading and takt for three candidate recipes before a drop of dye is used.
Agent, input and output
agent Yield-and-Takt result ← 3 candidates scored · best predicted ΔE 0.52 · 118 min cycle -
03 dye.predict_recipe SUCCEEDED 2.4s
Spectral recipe model proposes a 3-dye combination against the buyer standard, corrected for this substrate’s absorbency and the current water hardness.
Agent, input and output
agent Dye-and-Color result ← 2.14% Navy RGB · 0.61% Blue BRF · 0.08% Black B · salt 62 g/L -
04 human.approve_recipe APPROVAL 4m 12s
Recipe deviates >5% from the standing card, so autonomy is capped: the colourist approves, rejects or edits before any dosing. Guardrail policy MILL-02/DYE-APPROVAL.
Agent, input and output
agent Dye-and-Color result → awaiting colourist · approved on shift 04:18 -
05 dye.control_bath SUCCEEDED 118m
Closed-loop dosing and ramp control against live bath spectrophotometry; two mid-cycle corrections applied when exhaustion ran 4% ahead of model.
Agent, input and output
agent Dye-and-Color result ← exhaust 96.4% · 2 corrections · bath ΔE 0.41 at drain -
06 shade.verify_lot SUCCEEDED 9.1s
Per-roll spectral read against Indigo 19-4028 TCX under D65/TL84/A; metamerism scored and rolls sorted into shade bands.
Agent, input and output
agent Shade-and-Quality result ← mean ΔE 0.58 · 13 rolls band A · 1 roll band B -
07 finish.control_stenter SUCCEEDED 64m
Stenter chemistry, overfeed and curing profile set to hold hand, 148 cm width and residual shrinkage under 3%.
Agent, input and output
agent Finish-and-Hand result ← width 148.3 cm · shrink 2.1% warp / 1.4% weft -
08 inspect.scan_rolls FAILED 38m
In-line vision flags a recurring 6 cm weft streak on roll 09 between 412 m and 445 m, traced to a stenter pin-chain slip, not a dye fault.
Agent, input and output
agent Defect-and-Inspect result ← 1 defect cluster · roll 09 · 33 m affected · class WEFT_STREAK -
09 plan.rework_route SUCCEEDED 1.6s
cuOpt re-plans the cut and rework route: 33 m diverted to seconds, remaining 14,890 m re-sequenced so the buyer’s Thursday ship date still holds.
Agent, input and output
agent Yield-and-Takt result ← 33 m to seconds (0.22%) · ship date held -
10 quality.grade_and_release SUCCEEDED 3.0s
Four-point grading, roll genealogy and the conformance packet (AATCC 61, ISO 105-C06, ZDHC MRSL declaration) are assembled and released to the buyer portal.
Agent, input and output
agent Quality-and-Conformance result ← 13 rolls Grade A · 1 roll Grade B · packet signed
RUN OUTCOME
Lot released first time, on shade, on date.
| Measure | Value | Context |
|---|---|---|
| Final mean ΔE | 0.58 | against a 0.80 tolerance |
| Re-dyes avoided | 1 | the standing card would have missed shade |
| Seconds | 0.22% | versus the scenario mill baseline of 1.9% |
| Water used | 46 L/kg | versus 59 L/kg on the scenario recipe card |
| Human decisions | 1 | the recipe approval gate |
Why this matters
A mill does not fail on the good lots. It fails on the lot where the substrate drifted, the recipe card was wrong, and nobody found out until the buyer opened the carton. Yarneon makes that lot legible while it is still in the machine.
Reasoning trace
The reasoning, written down
Goal, observation, plan, action, decision, handoff, outcome. A colourist can read this and disagree with it, which is the point. Autonomy a mill cannot argue with is autonomy a mill will not adopt.
-
Land DL-4471 on Indigo 19-4028 TCX within ΔE 0.80, first time, before the Thursday 06:00 ship cut-off.
-
Greige absorbency for this lot reads 8% above the card baseline, and incoming water hardness is 142 ppm, both push the exhaust curve faster than the standing recipe assumes.
-
Simulate three recipes in the fabric twin before committing dye. Prefer the candidate with the lowest predicted ΔE that also fits the 118-minute machine slot.
-
twin.simulate(recipe_candidates=3, machine="jet-07")
-
Candidate B predicts ΔE 0.52 and finishes 6 minutes inside the slot. Candidates A and C either overshoot shade or miss takt.
-
Candidate B deviates 7.2% from the standing card, above the 5% autonomy ceiling for dye dosing. Escalate to the colourist rather than act.
-
Dye-and-Color → the colourist on shift, with the simulated curve, the deviation reason and a one-click revert attached.
-
plc.dose(machine="jet-07", schedule="ramp-B") after approval, with two mid-cycle corrections when exhaustion led the model by 4%.
-
Vision flags a 33 m weft streak on roll 09 after finishing. Correlating stenter telemetry shows a pin-chain slip at 04:51, a mechanical cause, not a dye fault.
-
Lot released at mean ΔE 0.58 with 0.22% seconds. A maintenance work order is raised against the stenter pin chain, and the streak signature is added to the mill’s defect memory.
Human in the loop
Craft is captured, not replaced
Textile mills run on scarce craft: the colourist who knows this jet runs hot, the weaving master who hears a beam going wrong. That knowledge is retiring faster than it is being replaced.
Yarneon writes it down. Every approval, override and correction becomes per-mill memory scoped to your tenant, so the mill keeps improving after the person who taught it has gone home.
- Per-colourist and per-operator performance memory
- Overrides captured as training signal, not noise
- Versioned memory, scoped per tenant, never shared across mills
Accelerated computing
Physical AI needs real compute, at the machine
Line-speed vision, spectral colour science and a fabric-line twin are GPU-essential workloads. Yarneon runs them at the mill edge and trains them centrally.
| Component | Role in the mill |
|---|---|
| Isaac + Jetson | Robotic roll, beam and doffing handling, plus mill-edge inference on inspection and control cells. |
| Metropolis / DeepStream + TensorRT | Line-speed fabric-defect vision pipelines targeting sub-100 ms per-frame decisions. ASPIRATIONAL |
| Holoscan | Sensor fusion across cameras, spectrophotometers, probes, PLC events and meters into one explainable state stream. |
| Omniverse + Replicator / Cosmos | The as-produced fabric twin, and synthetic generation of rare defect, barré and shade-variation cases. |
| cuOpt | Dye-lot sequencing, rework queues, machine assignment and water/energy allocation under hard constraints. |
| Triton + NIM + NeMo | Multi-model serving at the mill edge, plus grounded textile-process and colour reasoning with citations. |
| NVIDIA AI Enterprise | Private and on-prem deployment for groups protecting dye recipes and construction IP. |
Items marked ASPIRATIONAL describe target architecture ahead of full production validation.
Guardrails
An agent earns the right to touch a machine
Nothing here is a global "AI on" switch. Autonomy is granted per setpoint, bounded by policy, reversible by any operator and logged permanently.
- Graduated autonomy
- Every workflow moves observe → assist → bounded auto-control. Autonomy ceilings are per-setpoint policy, not a global switch.
- Human-in-the-loop gates
- Approval gates fire on deviation thresholds, new substrates, or any action outside the authorised envelope. Approvals are attributed and time-stamped.
- Reversibility
- Every control action ships with a revert path and a last-known-good setpoint. Operators can take manual control at any moment, from the HMI or the panel.
- Immutable audit log
- Assurance-grade, append-only record of every observation, decision, tool call, approval and actuation, exportable for buyer and regulatory audit.
- Tenant isolation
- Recipes, construction libraries and roll data are isolated per tenant. No cross-tenant training. Permission-aware retrieval with enforced citations.
- Sandboxed tools
- Agents call machines through a typed, scoped tool layer with rate and range limits. There is no path from a model output to an unbounded PLC write.
Operating figures
The numbers a mill manager asks about second
Latency targets, connector coverage, autonomy modes and the record: the practical questions after the headline.
<100 ms
Design target for a per-frame inspection decision on Jetson-class edge hardware. ASPIRATIONAL
8
Connector families: looms, frames, dyeing, colour, inspection, MES, utilities, robotics.
3
Autonomy modes: observe, assist, bounded control. Granted per setpoint.
1
Record per mill: append-only, exportable, shared with buyers on request.
Items marked ASPIRATIONAL are design targets ahead of production validation.
Questions
Questions about releases and control
Both, in that order. Every workflow starts in observe mode, graduates to assist (the agent proposes, a human commits), and only then to bounded auto-control on the specific setpoints you authorise. Autonomy ceilings, approval gates and spend limits are policy objects you own, and every action is reversible and logged.
The control loop lives at the mill edge. Inference, guardrails and the audit log run locally on the edge appliance, so a lost uplink degrades reporting and retraining, not machine control. Buffered telemetry syncs when the link returns.
Design partners start with one wedge line in shadow mode: connectors and telemetry first, then models scored against what the mill actually produced, then assist mode where the agent proposes and an engineer approves. Whole-mill rollout follows once the wedge holds its numbers. Durations depend on how instrumented the mill already is, so we quote them after the assessment, not before.
To predict recipes well, yes, and that IP never leaves your tenant. Retrieval is multi-tenant isolated, models fine-tuned on your data are yours, and on-prem deployment is available for groups that will not put recipes in any cloud.
A Jetson-class edge device per inspection or control cell, one GPU server per mill for multi-model serving, line-scan cameras where in-line inspection is in scope, and a spectrophotometer feed for colour work. We survey the existing fleet first, most mills already have more sensing than they use.
Want to be notified when the first release lands?
Write to us and we will tell you when something ships, and what it is allowed to touch.
