Skip to Main Content
Leak Sight AI
Product
WorkflowIncidentCapabilitiesArchitecture
Pilot
Company
AboutContact
Request a Pilot
ProductWorkflowIncidentCapabilitiesArchitecture
PilotPilot Program
CompanyAboutContact
Request a Pilot
Leak Sight AI / Legal

Current notice

Product Disclaimer

The safety, operational, model-output, and product-information boundaries for Leak Sight AI website material.

Effective date
7 August 2026
Applies to
Website and product information
Document statusCurrent notice

Leak Sight AI is decision support for trained industrial teams. It is not an emergency service, autonomous safety controller, or substitute for approved facility procedures.

On this page
  1. 1. Decision Support Only
  2. 2. Not an Emergency Service
  3. 3. Detection Boundaries
  4. 4. Model Outputs & Confidence
  5. 5. Website Material
  6. 6. Technology & Integration References
  7. 7. Facility Responsibility
  8. 8. Questions

1. Decision Support Only

Leak Sight AI helps trained personnel review potential visible water and steam leaks, supporting evidence, probable location, estimated urgency, and a suggested next action. Human approval remains required for critical actions.

Do not use a Leak Sight AI output as the sole basis for isolation, shutdown, lockout, evacuation, equipment restart, regulatory reporting, or a decision that affects worker or public safety.

2. Not an Emergency Service

This website does not monitor live facilities, receive active alarms, dispatch responders, or provide emergency assistance. Follow the facility’s approved emergency, maintenance, cyber, and incident-response procedures.

3. Detection Boundaries

The supported product scope covers visible water and steam around exposed above-ground pipes, valves, joints, pumps, and related equipment. Buried, occluded, small, intermittent, invisible, or out-of-view leaks may not be observable by supported cameras.

Weather, washdown, condensation, shadows, glare, exhaust, damaged insulation, changing camera views, network conditions, sensor quality, and facility operations can produce missed events or false alarms. Calibration and temporal review reduce uncertainty but do not eliminate it.

4. Model Outputs & Confidence

Confidence values, severity labels, probable sources, and recommended actions are model or workflow outputs, not statements of certainty. They must be reviewed with current operating conditions, asset history, sensor context, and approved procedures.

5. Website Material

Interface views, confidence values, timelines, evidence traces, diagrams, and photographs communicate product functionality. They are not detection-rate claims, savings estimates, endorsements, or field-performance guarantees.

6. Technology & Integration References

Leak Sight AI supports NVIDIA Jetson, TensorRT, and DeepStream for edge inference. Maintenance integrations are defined for each facility. Third-party references do not imply endorsement, partnership, or certification.

7. Facility Responsibility

Each organization remains responsible for camera placement, inspections, maintenance, engineering decisions, worker notification, privacy, cybersecurity, data retention, regulatory compliance, and safe operation of its assets.

A signed pilot or service agreement may define additional validation, support, performance, security, and liability terms for a specific deployment.

8. Questions

Use the Contact page for questions about product scope or this disclaimer. Do not send live incident reports or confidential facility information through the public form.

All policiesOpen the Legal Center
Related documentWebsite Terms
Related documentAcceptable Use Policy
Questions or requestsContact Leak Sight AI
Leak Sight AI

Multimodal leak intelligence for exposed industrial water and steam systems.

ProductWorkflowIncidentCapabilitiesArchitecture
PilotProgram StructureRequest a Pilot
CompanyAboutContact
LegalLegal CenterPrivacy PolicyTerms
© 2026 Leak Sight AIDecision support only. Human approval is required for critical actions.