░ trace runtime v2.0 · safety-first ░

Diagnose any plant fault —
without crossing a safety boundary.

TRACE is a decision-support engine for PLC and SCADA equipment in steel, cement, mining and process plants. It cross-references ladder logic, P&ID diagrams, calibration history and a compounding Fault Pattern Database — then refuses to guess across LOTO, SIS or ATEX boundaries. Operators get plain language. Engineers get full depth. Nobody crosses a hazard zone without authorisation.

trace://session/live active
machineCaster Line A · CCM-3200
hazardATEX Zone 2 · SIS present
intakesymptom + observable behaviour
matchFPD #ddddeeee · conf 88% · ×7 prior
hypothesis · 88

Mold oscillation servo thermal overload — cooling filter F-22 clogged. Overdue calibration on TT-204 contributing.

fix · 12 min · LOTO required
  1. Confirm LOTO on drive cabinet D-4
  2. Replace cooling filter cartridge F-22
  3. Engineer to validate amplitude 4–6 mm
verify · plausibletrace v2.0
protocol

Six-step intake. Four safety gates. Zero guessing.

01

Session

Emergency, scheduled, or performance check.

02

Machine

Confirm running PLC version against record.

03

Calibration

Overdue sensors surfaced before diagnosis.

04

Intake

Error code, symptom, quality, performance or physical.

05

State

Tag/value pairs of current operating conditions.

06

Confirm

Other active faults? TRACE runs with full context.

▸ safety gate
SIS boundary

Safety Instrumented System faults are refused — escalated, not diagnosed.

▸ safety gate
LOTO gate

Every physical fix requires explicit lockout/tagout confirmation.

▸ safety gate
ATEX block

Hazardous atmospheres force engineer-validated procedures.

▸ safety gate
Version flag

PLC code mismatch reduces confidence and warns the engineer.

compounding intelligence

Every confirmed fix becomes future leverage.

When an operator confirms a resolution, TRACE upserts it into the Fault Pattern Database with confidence weighting, hazard flags and source-session lineage. The next time the same signature appears — on any machine of the same model, at any client — diagnosis is faster, surer, and already pre-validated against your safety envelope.

  • Skill-aware output — basic operators get plain language, engineers get ladder/ST depth.
  • Per-client RLS — Acme can never see Pioneer's machines, documents, calibration or patterns.
  • Immutable audit log — every diagnosis, escalation and resolution is append-only.
  • Engineer console — escalated, blocked and SIS-bounded sessions queued for review.
fault_patterns · live extract
signaturemodelconfseen
E-1042CCM-320088%×7
E-2207RF-88082%×4
E-555KD-50091%×11
E-9001CV-70076%×3
workflow

From alarm to verified fix in three surfaces.

01 · intake

Six-step guided intake

[ machine: CCM-3200      ]
[ intake: SYMPTOM         ]
[ describe: vibration trip ]
[ state:  speed=82% temp=91 ]

Operator picks the session type, confirms running PLC version, then describes the symptom in plain language or pastes an error code.

02 · diagnose

Ranked causes, gated fixes

► cause 1 · cooling F-22  88%
► cause 2 · servo bearing 41%
▸ fix    · 12 min · LOTO
▸ verify · plausible       

TRACE returns confidence-weighted root causes against PLC code excerpts, P&IDs and the live Fault Pattern Database. SIS, LOTO and ATEX gates fire before any fix renders.

03 · escalate

Engineer console, in-app alerts

▣ alerts (2)              
  escalation · URGENT     
  filter cartridge clog   
  → engineer notes ⤴      

Blocked, SIS-bounded or requires-engineer sessions queue for the on-call engineer with a realtime in-app alert. Resolution notes flow back to the operator the same way.

stack

Engineered to deploy in real plants, today.

Auth
Email + per-client RLS
AI Core
Gemini 2.5 Pro · structured · temp 0
Safety
SIS / LOTO / ATEX post-processor
Documents
PDF page-by-page extraction · 9 types
DB
Postgres · RLS · immutable audit
Runtime
TanStack Start · edge SSR
Realtime
In-app alerts on escalation
Health
/api/public/health · live checks
faq

Answers for engineers, not marketers.

Everything plant managers, controls engineers and SI partners ask before deploying TRACE on a live line.

still unsure?

Book a 20-minute walkthrough against your own PLC code excerpt and a sample fault.

request a walkthrough →

░ ready to deploy ░

Your next unplanned downtime
is a solved problem.

Spin up TRACE for your fleet. Seeded demo clients, machines and fault patterns are already loaded — log in and run a real diagnosis in under a minute.