Irisity provides AI-powered video analytics for security monitoring, forensic investigation, operational analysis, and selected health and safety applications. Its IRIS+ platform analyzes compatible video streams, creates searchable metadata, applies configured detection rules, and generates events when defined activity appears within a camera view.
Organizations evaluating video analytics should define the events they need to identify, how operators will respond, and which systems will receive the resulting information. NIST’s work on interoperable video analytics emphasizes governance and system integration, while CISA’s security and resiliency guide provides a broader framework for assessing facility security and resilience requirements.
Irisity is a Gothenburg-based video-analytics company founded in 2006. Its product history includes remote guarding, perimeter monitoring, central monitoring, forensic investigation, and enterprise video intelligence.
The company expanded its portfolio through the acquisition of Agent Video Intelligence. Its current environment includes IRIS+ products for enterprise surveillance, professional video investigation, and central monitoring, alongside supported legacy products.
Published applications span:
Security and surveillance remain the platform’s principal focus. Health, safety, traffic, and operational analytics are available within parts of the wider product environment, depending on the selected edition, modules, and configuration.
IRIS+ receives video from compatible cameras and processes it through edge devices, local servers, or other configured infrastructure. The platform extracts information about visible people, vehicles, objects, movement, and defined events.
Its architecture includes:
IRIS+ Core can be hosted in the cloud or deployed within an environment selected by the customer. Air-gapped configurations are also available for sites that cannot maintain an external internet connection.
The management environment allows users to configure camera connections, event rules, alert delivery, permissions, analytics, and integrations. Health-monitoring functions can also identify issues involving video quality, interrupted streams, blocked views, connectivity, and processing components.
IRIS+ allows users to define areas, lines, object types, behaviors, schedules, and conditions that determine when an event should be generated.
Published analytics include:
These capabilities can help security operators prioritize footage across camera networks that are too large to observe continuously.
Performance depends on the physical and technical environment. Relevant factors include:
A detector that performs adequately in one camera view may require different configuration or positioning in another.
IRIS+ creates metadata that can be searched after an incident or operational event. Available investigation functions may include:
IRIS+ Professional supports both historical queries and live rules. A query used to search recorded footage may therefore also be configured to create an alert when the same type of event occurs.
Some newer search functions are described as beta features. Buyers should test the supported language, search reliability, processing requirements, and relevance of returned results before including them in an investigation procedure.
Current IRIS+ Professional documentation includes several safety-related queries, including:
The near-miss query evaluates whether a moving object comes within a configured distance of a person during a specified time interval. Users can adjust variables involving confidence, movement, sampling, distance, time, and the monitored region.
This may support selected vehicle-pedestrian monitoring scenarios. It does not establish that the same configuration is suitable for every forklift, pallet jack, yard tractor, crane, truck, or industrial near miss.
Current documented PPE queries focus primarily on helmets and safety vests. Organizations requiring detection of goggles, gloves, respirators, hearing protection, harnesses, bump caps, or other equipment should confirm whether those classes are available through standard models or require additional configuration.
Irisity offers personalized models for objects, attributes, and events that are not covered by standard analytics. A customer may provide representative examples of the item or condition that needs to be recognized.
Potential applications include:
Personalized models can extend the platform’s coverage but also add project requirements. These may include collecting representative footage, labeling examples, validating results, adjusting camera views, monitoring model performance, and retraining the model when site conditions change.
Buyers should clarify who performs this work, how model changes are approved, whether additional licensing applies, and how long-term maintenance is handled.
IRIS+ supports ONVIF and RTSP video streams from fixed, angled, overhead, optical, and supported thermal cameras.
Integration options include:
Integration depth varies by system. Some connections deliver alerts into the video-management platform, while others provide camera access, metadata, health information, or investigation functions.
Organizations should confirm compatibility with their exact camera models, firmware, recording architecture, VMS version, authentication method, and network design.
Irisity has operated in video analytics for approximately two decades. Its product history covers remote guarding, alarm filtering, perimeter monitoring, central monitoring, forensic analysis, and enterprise surveillance.
This experience may be relevant to security organizations managing outdoor environments, distributed camera networks, established monitoring centers, or a range of conventional surveillance events.
IRIS+ supports:
This range gives organizations options for balancing latency, bandwidth, data residency, cybersecurity, infrastructure ownership, and operating support.
Architecture selection still requires technical planning. Edge systems need sufficient access to camera streams, while cloud and hybrid designs require appropriate connectivity, security controls, and operational monitoring.
Irisity offers floating licenses that can be reassigned between supported cameras rather than remaining permanently attached to a single view.
This may be useful when:
Buyers should confirm which editions and analytics support license movement and whether reassignment affects configuration, historical information, or event continuity.
IRIS+ combines live alerts, camera management, system-health information, and forensic search within a centralized environment.
Security personnel can use the platform to review active events and investigate historical footage without relying entirely on manual timeline searches. System-health monitoring may also help distinguish between an absence of alerts and a technical issue preventing analysis.
Irisity supports PPE, near-miss, restricted-area, fall, fire, and other safety-related applications. However, its broader product positioning remains centered on surveillance, security operations, central monitoring, and investigation.
Industrial safety teams should confirm whether the proposed configuration supports each required risk. Relevant questions include:
Some requirements may need personalized models, additional systems, or separate engineering controls.
IRIS+ allows users to adjust confidence, sensitivity, distance, sampling, movement, and timing thresholds. These controls provide flexibility but also make performance dependent on site-specific configuration.
A pilot should include:
Results from one analytic, such as perimeter intrusion, should not be assumed to represent performance for a different application such as PPE or vehicle-person proximity.
Irisity publishes platform documentation, integrations, customer statements, and security use cases. Detailed public industrial case studies containing named facilities, baseline safety measures, implementation periods, event reductions, and financial outcomes are less common.
Independent public user reviews are also limited. This makes it more difficult to assess day-to-day administration, support, alert quality, model maintenance, and upgrade experiences from publicly available end-user evidence alone.
Prospective customers should request:
Personalized analytics can extend the platform, but a custom model should not be treated as equivalent to an established standard capability.
A custom project may require:
This work should be included when evaluating schedule, cost, internal staffing, and support responsibilities.
IRIS+ can generate alerts, searches, metadata, statistics, and integrations. Public materials place less emphasis on complete EHS corrective-action management.
Safety teams should confirm how an identified event becomes:
These functions may need to be managed through an EHS platform, work-management system, video-management workflow, or custom integration.
IRIS+ supports facial recognition and other identity-oriented investigation functions in parts of its product environment. Irisity also provides anonymization options.
Organizations considering biometric analytics should review:
Facial recognition, anonymized monitoring, and general forensic investigation may require different policies even when they operate through the same platform.
Irisity may be relevant when an organization needs several of the following:
The platform may be particularly relevant to security-led projects involving an established camera and video-management environment.
Industrial facilities evaluating IRIS+ for workplace safety should define which functions will remain within the surveillance environment, which will be managed through an EHS system, and which risks require separate sensors, engineered controls, or safety equipment.
Voxel focuses on industrial safety and operational risk rather than broad surveillance and biometric investigation.
Its published applications include supported events involving:
Irisity may be relevant when the primary objective is security monitoring, investigation, or centralized video analytics. Voxel may be relevant when the main objective is identifying recurring industrial risks and organizing the safety response.
Neither platform replaces machine guarding, emergency systems, vehicle warning equipment, required inspections, or worker participation.
Voxel’s AI for risk management connects supported detections with video evidence, trends, reporting, and follow-up workflows.
Teams can:
These functions address the process that follows a detected industrial-risk event. They do not replace Irisity’s broader surveillance, facial-recognition, forensic-search, and central-monitoring capabilities.
The Port of Virginia results provide an example of camera-based industrial monitoring within a complex transportation environment.
The port used Voxel to review truck speeding, failure-to-stop events, PPE practices, and activity in high-traffic areas. Voxel adapted a vehicle-monitoring algorithm to assess trucks and reduced the amount of time the safety team spent manually reviewing recorded video.
Within six months, the port reported:
Daily footage-review time decreased from approximately two to three hours to 20 to 30 minutes. The reported outcomes reflect the port’s use of the platform alongside communication, enforcement, hazard removal, and other management actions.
Voxel does not use facial recognition. Its published privacy and security controls include:
These controls support a model centered on site-related events rather than facial identification. Organizations should still establish policies covering camera scope, access, retention, worker communication, coaching, and permitted use.
The appropriate platform depends on the intended program. Irisity may align with security-led surveillance and investigation, while Voxel may align with EHS-led industrial-risk monitoring and corrective-action workflows.
Organizations can contact Voxel directly to review supported industrial events, existing camera coverage, privacy requirements, and action-management needs.
IRIS+ supports events such as intrusion, loitering, line crossing, unattended objects, smoke, fire, crowds, vehicles, and restricted-area activity. Some editions also include PPE, near-miss, facial-recognition, license-plate, and natural-language investigation functions. Available analytics depend on the product edition, licensed modules, and configuration. Buyers should confirm the current feature set for their intended deployment.
Irisity supports selected safety applications, including helmet and safety-vest detection, vehicle-person near-miss analysis, falls, fires, and restricted areas. These functions may contribute to safety monitoring when configured and validated for the site. Public documentation does not establish coverage for every industrial risk, vehicle type, or ergonomic condition. Facilities should test each required use case under representative operating conditions.
IRIS+ supports ONVIF and RTSP streams from many fixed, angled, overhead, optical, and supported thermal cameras. Existing cameras still need suitable resolution, frame rate, angle, object visibility, and network connectivity for the selected analytic. Protocol compatibility alone does not guarantee adequate detection performance. A technical camera assessment should be included in the evaluation.
Irisity primarily focuses on broad surveillance, real-time alarms, forensic investigation, and multi-industry video analytics. Voxel focuses on supported camera-visible industrial risks and workflows for coaching, corrective actions, trends, and cross-site safety reporting. Irisity includes selected workplace-safety analytics, while Voxel does not provide the same range of biometric and general-security investigation functions. Selection should reflect whether the primary need involves surveillance, industrial safety, or separate systems for both.
A pilot should include representative cameras, lighting, weather, vehicles, PPE, worker activity, and normal operating conditions. Teams should measure missed events, false alerts, alert timing, configuration effort, hardware load, bandwidth, and operator workload. They should also test VMS integration, permissions, anonymization, retention, investigation, and escalation procedures. Expansion should follow demonstrated performance for the specific analytics and camera environments being evaluated.
Explore AI Summary
