How a piece of software is allowed to touch a line.
Chassira writes to production equipment, so the security story is mostly a control story: a bounded action space, a policy engine before every write, locked safety-critical parameters, named approvers, an append-only audit log the plant owns, hard fallback to the controller and its safety-rated PLCs, and isolation between tenants and their programs.
- Source evidence it acted on
- Specs, control plans and standards it retrieved, with citations
- The approved envelope and the limit applied
- Confidence and estimated quality, rework, downtime and warranty impact
- Rollback path
- Responsible approver
- Body and vehicle scope
Nothing is written without this record.
An agent near a crash-safety joint needs a leash
Chassira writes to production equipment. Every capability is scoped, every write is policy-checked against the qualified envelope, safety-critical parameters are locked, and every action is written to an append-only audit log the plant owns.
- Bounded action space. Each agent writes only to an explicit tag allow-list, inside per-tag rate and magnitude limits and the qualified process envelope for the joint family, colour block or vehicle program.
- Policy engine before every write. Autonomy level, shift, program, interlock state and operator presence are evaluated before a setpoint moves. NeMo Guardrails enforces spec, IATF, functional-safety, cyber and human-approval policy before any action writes back to robots, MES, QA or work instructions.
- Safety-critical parameters locked. Weld schedules wait for a named body-shop engineer until the joint family is qualified. Torque specifications, battery-joint parameters and end-of-line test limits and verdicts are read-only to every agent, always.
- Immutable audit log. Append-only and exportable. Every action records source evidence, the approved envelope, confidence, estimated quality, rework, downtime and warranty impact, rollback path, approver and vehicle scope.
- Hard fallback. Loss of the edge node, the network or a model returns control to the press, robot, booth or MES controller's last known-good state. Safety-rated PLCs and interlocks are never in Chassira's path.
- Tenant and program isolation. Vehicle designs, control plans, specs, models and memory never cross an OEM boundary. On-prem and plant-edge deployment are available for regulated plants.
Compliance posture
| Item | Scope | Status |
|---|---|---|
| SOC 2 Type I | Cloud control plane | QUEUED Planned, first 6 months, with an IATF and functional-safety-aware review |
| SOC 2 Type II | Cloud control plane | QUEUED Planned, 6 to 12 months |
| SSO and RBAC | Console and approvals | QUEUED Planned, 6 to 12 months |
| IATF 16949 and APQP assurance packages | Disposition and traceability evidence | QUEUED Planned, 6 to 12 months; not a certification claim |
| ISO 26262-aware control | Functional-safety requirements cited in reasoning and gates | RUNNING In the design; not a certification claim |
| Standards-cited reasoning | IATF 16949, APQP and PPAP, torque specs, recall procedures | RUNNING In the design, every answer cites its source |
| Per-tenant and program isolation | Designs, control plans, models, memory | RUNNING In the design from day one |
| On-prem and plant-edge deployment | Regulated OEM plants | RUNNING Design option |
What data exists and who can see it
The data strategy from the architecture: body- and vehicle-based, labelled from the floor, synthetic only for rare defects, governed per tenant.
Sources
- Press force and stroke, blank feed, die condition, panel dimension
- Weld current, force and expulsion traces, seam vision, rivet and adhesive signals, in-line metrology, gap-and-flush images
- Booth environment, film build, oven profiles, paint-defect images
- Torque traces, joining and integrity results, fill volumes, end-of-line and ADAS test results, non-conformance records and the as-built record
Governance
- Tenant isolation for vehicle designs, control plans, specs, models and memory
- Program IP redaction and data-rights controls
- Model cards and dataset versioning
- Traceable consent before any cross-plant learning, which is planned and not yet in place
- Synthetic data used only to augment rare defects, validated against real distributions
The write path
What has to be true before a setpoint moves.
- A declared tool. The agent can only call tools declared with tags, units, limits and a qualified envelope. Torque specifications, battery-joint parameters and test limits cannot be declared writable.
- A policy decision. Autonomy level, shift, program, interlock state and operator presence are evaluated. NeMo Guardrails enforces spec, IATF, functional-safety and approval policy.
- A simulated check. Where the agent is configured with verify=twin, the move is simulated on the plant twin before it is written.
- An approval, if required. Weld-schedule changes, torque-related decisions and vehicle release wait for the named approver.
- A bounded write. The connector enforces per-tag rate and magnitude limits again at the OT boundary. Safety-rated PLCs and interlocks are never in Chassira's path.
- A record. The action, its evidence and its outcome are appended to the audit log and the vehicle's as-built record.
- A rollback path. Declared rollback conditions return the tag to its previous value or the controller to last known-good.
Four levels, set per agent and per tag
A car plant does not go from manual to unattended in one step. Chassira makes the level explicit, auditable and reversible at any time. The first release runs shadow mode, then assist mode, then bounded autonomy on low-risk loops. Weld schedules stay supervised until the joint family is qualified, and torque specifications and test verdicts are never written by an agent.
| Level | What the agent does | What the human does | Where it sits in a pilot |
|---|---|---|---|
| L1 · Shadow | Reads weld, metrology, vision, torque, test, robot and MES signals; predicts and recommends; writes nothing | Runs the line as today; compares predictions to audit and engineer outcomes | Design partners start here; the baseline is measured from the plant's own records |
| L2 · Assist | Proposes weld, framing, paint, routing, holding and sequencing moves with evidence, confidence, envelope and rollback path | Approves each action in the review console | Assist mode: accuracy and ROI proven against the signed baseline |
| L3 · Bounded | Executes pre-approved low-risk actions: booth environment, framing offsets, blank feed and lubrication, routing, holds, sequencing, inside PLC and safety limits and the qualified envelope | Approves weld-schedule changes, torque-related decisions and vehicle release | Graduated autonomy, after audited accuracy and ROI |
| L4 · Scaled | Supervised weld-schedule adaptation for qualified joint families, across lines and plants | Sets the envelope; reviews the shift record and evaluation results | Planned; after golden-dataset evaluation, twin validation and quality sign-off. Torque specifications and test verdicts never leave human hands |
One scenario shift, step by step
Every Chassira run is an ordered, inspectable sequence. This is run_41a9c2 on Plant BIW-2 · EV crossover body shop to end of line · shift 1. Scenario completes in 10 steps with one approval gate. Weld schedule change on robot W-14 approved by the body-shop launch engineer, framing offset held inside the envelope, booth humidity trimmed before the predicted dirt defects appeared, one pack held at marriage on a torque trace, and the shift released with full as-built records.
-
01
ingest.shift
SUCCEEDED
0.8 s
Plant Orchestrator · Pulled the shift build sequence, program control plan and qualified envelopes from MES: 412 bodies, four trim variants, weld schedule set W-rev 7, framing targets per station, paint colour blocks, torque specifications. Locked every safety-critical parameter as read-only.
-
02
twin.simulate
SUCCEEDED
1 m 40 s
Vehicle-and-Line Twin · Simulated 18 candidate weld-sequence, framing-offset and booth-window combinations on the plant twin, ranked on predicted body geometry, expulsion risk at robot W-14, paint defect map and takt against launch cadence.
-
03
stamp.form
SUCCEEDED
continuous
Stamp-and-Form · Blank feed on press line P-3 adjusted by 0.4 mm and lubricant dose raised 3 percent for the new coil lot as springback risk on the rear door inner climbed. Panel dimension held inside the window; genealogy handed to framing.
-
04
approve.human
APPROVAL
3 m 12 s
Plant Orchestrator · Candidate #6 lowers weld current 2 percent and adds 20 ms hold on robot W-14 to cut predicted expulsion. Weld schedules are crash-safety critical and stay with a human, so the change was held for the body-shop launch engineer and approved at 06:14:52.
-
05
weld.join
SUCCEEDED
continuous
Weld-and-Join · Approved schedule 7b applied on W-14. Current and force traces confirmed inside the window on the first 40 bodies; predicted expulsion risk fell from 0.31 to 0.06 and no geometry debt was introduced at the adjacent seams.
-
06
frame.geometry
SUCCEEDED
continuous
Frame-and-Geometry · Framing station FS-1 offset corrected by 0.3 mm on the left B-pillar datum inside the qualified envelope as in-line measurement trended toward the tolerance edge. Gap-and-flush vision confirmed the correction on the next 25 bodies.
-
07
paint.finish
SUCCEEDED
continuous
Paint-and-Finish · Booth B-2 humidity trimmed 2 points and airflow rebalanced for the dark-colour block as the defect model predicted dirt on horizontal panels. Film build held inside the window on the first colour block.
-
08
marriage.torque
SUCCEEDED
continuous
Powertrain-and-Battery · Pack PK-1182 held at marriage after a torque-to-yield trace on fastener 7 fell outside the specification curve. Routed to rework with the trace attached; torque specification untouched, as always.
-
09
inspect.fuse
SUCCEEDED
continuous
Quality-and-Anomaly · Gap-and-flush and paint-defect cameras, weld and torque traces and test results fused per body. Two paint dirt defects localised to booth B-2 zone 3 before the humidity trim took effect; both routed to spot repair, none escaped to final.
-
10
eol.release
SUCCEEDED
2 m 05 s
EOL-and-Calibrate · End of line proved alignment, ADAS calibration, leak, brake, roll and functional tests on every vehicle; two ADAS retests sequenced inside takt. Shift 1 released with as-built records, approvals and dispositions written for the quality lead.
What the audit log looks like
Every tool invocation, argument and result is written to an append-only, human-readable log, and every reasoning step is expandable. Nothing about a run is hidden from the plant.
- 06:00:02mes.read_sequence({ plant: "BIW-2", shift: 1, program: "EV-CUV" })
- 06:00:03ok 412 bodies · 4 variants · weld set W-rev 7 · torque specs read-only
- 06:00:04twin.simulate({ candidates: 18, horizon: "1 shift" })
- 06:01:44ok best #6 · geometry inside targets · expulsion risk 0.06 · takt 61 s
- 06:01:45policy.evaluate({ candidate: 6, autonomy: "L2+" })
- 06:01:45ok 6 writes permitted · weld_schedule requires human approval · torque_spec locked
- 06:01:46blank.set_feed({ press: "P-3", delta: "+0.4 mm", lube: "+3%" })
- 06:02:30ok springback risk 0.42 → 0.11 · panel dimension in window
- 06:02:31approval.request({ action: "weld_schedule", robot: "W-14", current: "-2%", hold: "+20 ms" })
- 06:14:52ok approved by body-shop launch engineer · logged
- 06:14:53weld.apply_schedule({ robot: "W-14", rev: "7b" })
- 06:48:10ok 40 bodies in window · expulsion risk 0.31 → 0.06
- 06:48:11frame.set_offset({ station: "FS-1", datum: "B-pillar L", delta: "-0.3 mm" })
- 07:12:44ok 25 bodies inside targets · gap-and-flush confirmed
- 07:12:45booth.set_env({ booth: "B-2", humidity: "-2 pt", airflow: "rebalance" })
- 07:40:02precursor: dirt risk 0.58 · booth B-2 zone 3 · dark colour block
- 07:40:03ok 2 defects localised · routed to spot repair · 0 escaped
- 08:05:19unit.hold({ pack: "PK-1182", fastener: 7, reason: "torque trace off curve" })
- 08:05:19ok held at marriage · routed to rework · spec untouched
- 08:05:20sensors.fuse({ cams: 24, traces: true, tests: true })
- 13:40:00retest.sequence({ vehicles: 2, test: "ADAS calibration" })
- 13:52:31ok 2 retests passed inside takt
- 13:58:00mes.log_vehicle({ shift: 1, records: "as-built" })
- 14:00:05ok scenario complete · 10 steps · 0 escapes · shift released
Reasoning trace
Plan, thought, action, observation, exactly as the orchestrator would record it.
GoalRun shift 1 on plant BIW-2 inside the qualified envelopes, hol…
Run shift 1 on plant BIW-2 inside the qualified envelopes, hold body geometry at the framing station, keep expulsion and paint defects from escaping, hold every torque to specification and prove every vehicle at end of line.
PlanSimulate weld-sequence, framing and booth candidates on the tw…
Simulate weld-sequence, framing and booth candidates on the twin first. Lead with stamping so panels arrive in dimension, then propose the weld schedule and hold it for a human, then correct framing inside the envelope, then set the booth ahead of the colour block. Fuse inspection continuously. Never touch a torque specification or a test verdict.
ThoughtThe standing weld schedule on W-14 runs current high enough to…
The standing weld schedule on W-14 runs current high enough to guarantee the joint on the thickest stack. With the new coil lot and panel dimension held at P-3, the twin predicts the same joint quality at 2 percent lower current with a 20 ms hold, and lower expulsion, because the panels arrive flatter.
Actiontwin.simulate(candidates=18) returns #6: blank feed +0.4 mm on…
twin.simulate(candidates=18) returns #6: blank feed +0.4 mm on P-3, weld current -2 percent and +20 ms hold on W-14, framing offset -0.3 mm at FS-1, booth B-2 humidity -2 points for the dark block.
ObservationCandidate #6 predicts geometry inside targets, expulsion risk …
Candidate #6 predicts geometry inside targets, expulsion risk 0.06, a lower dirt map and takt at 61 seconds. Three candidates scored better on takt but pushed predicted B-pillar variation past the tolerance edge and were discarded.
ActionExecute #6: six bounded writes permitted on stamping, framing,…
Execute #6: six bounded writes permitted on stamping, framing, booth, routing and sequencing; the weld-schedule write routed to the body-shop launch engineer; torque specifications and test limits locked.
ObservationDirt precursor in booth B-2 zone 3 at 07:40:02 before the humi…
Dirt precursor in booth B-2 zone 3 at 07:40:02 before the humidity trim took effect. Two defects localised and routed to spot repair; none escaped. One pack held at marriage on a torque trace at 08:05:19.
OutcomeScenario complete. Geometry inside targets, no expulsion escap…
Scenario complete. Geometry inside targets, no expulsion escapes, paint defects caught in station, one pack reworked, two ADAS retests passed inside takt. Shift 1 released with full as-built records.
Standardise autonomy across the fleet
One policy model, one audit trail, one benchmark across every process in every plant, with the weld, paint, geometry and battery models kept private to each plant and program.
- SSO and role-based access down to the tag level, planned for the 6 to 12 month release
- Fleet-wide autonomy policy with per-plant override and approval chains
- Cross-plant benchmarking on first-time-through, rework, repaint, downtime and warranty
- On-prem and plant-edge deployment for regulated OEM plants
- Edge autonomy that survives control-plane loss by returning to the controller's last known-good state
- Dedicated deployment engineer and quarterly model review per plant
Straight answers
The questions plant managers, launch engineers and quality leads ask in the first meeting.
Yes, but only within an explicit tag allow-list with per-tag rate and magnitude limits, inside the qualified envelope, and only at the autonomy level your plant has set. Design partners start in shadow mode, where Chassira writes nothing. Weld-schedule changes wait for a named body-shop engineer, and torque specifications, battery-joint parameters and end-of-line test limits and verdicts are read-only to every agent, always.
Control returns to the press, robot, booth or MES controller's last known-good state. Chassira is a supervisory layer on top of the safety-rated PLCs and interlocks you already run, never a replacement for them, so an outage degrades the line to its current way of running, not to a stop.
Advisory recommendations start from the plant twin and the control plan in the first weeks. Weld, geometry, paint, torque and end-of-line prediction improve as line history and labelled outcomes accumulate per program. The compute plan sizes onboarding at 150 to 600 GPU-hours per plant and 50,000 to 250,000 synthetic variants per program, all design targets.
Only if you choose cloud training. Designs, control plans, specs, process models and quality records are isolated per tenant and never used to train another OEM's models. On-prem and plant-edge deployment are available for regulated plants, and federated cross-plant learning is planned and waits on data-rights agreements.
You are, the same as with any control strategy, which is why every write is policy-checked, bounded, logged and reversible, and why anything crash-safety or IATF relevant waits for a named approver. The audit log records the request, the evidence, the envelope applied and the human decision.
A paid pilot on one process or area at the Process tier, scoped to a single wedge with a success metric written into the order form: weld or geometry rework, downtime minutes, paint defect or repaint rate, warranty or PPM, or launch quality. Shadow mode first, then assist mode with your engineer approving each move, then bounded autonomy on low-risk loops once ROI and safety gates are met.
Send us your security questionnaire
We answer plainly, including where an item is planned rather than done.