IREX is an Ethical AI and big-data platform for city governments, public-safety agencies and transportation authorities. We connect the cameras and sensors a city already owns to a private cloud it controls, analyze them in real time, and give people the chance to prevent an incident instead of reconstructing it afterwards.
Our Mission
IrexAI Inc. was incorporated in Delaware on 21 November 2019 and is headquartered in Murrieta, California. What began as a computer-vision problem became a governance one: the technology to watch a city has existed for years; the accountability to deserve that power has not.
So we built the platform the other way round. Constraints first, audit trail first, customer ownership first, and then the analytics on top. Today IREX runs on 250,000+ cameras across eight countries, and every high-risk action any operator takes on any of them is bound to a documented lawful reason.
The most powerful AI must also be the most accountable. Our logs are open to audit. Our constraints are verifiable. Our commitment is permanent.
What We Hold To
We bring academic rigor to the complexities of city governance, and build for the transparency of law enforcement that trust between citizens and police depends on. We work with non-profits to find missing children and adults, and with schools to curb drug distribution.
IREX identifies only pre-registered people and vehicles: suspects, missing persons, vehicles flagged in a case. It does not identify or track random individuals in public spaces. Customers hold 100% ownership and control of their data.
We run background checks on every prospective client and project to ensure fair use of the platform, and actively decline business with parties we judge high-risk in relation to data collection and security.
Ethical AI
Rather than treat ethics as a compliance checkbox, IREX embeds accountability in the architecture. Six pillars of trust span full transparency, cybersecurity, narrow constraints on law enforcement, permission-driven access, bias mitigation and regulatory compliance, plus the Case ID mandate: a world first that binds every high-risk AI action to its documented lawful reason.
Explore Ethical AI: the Six Pillars and Case ID →Data Sovereignty
A public-safety agency cannot outsource custody of the evidence it is accountable for. IREX is user-hosted by design: the platform installs on infrastructure the customer owns, whether bare metal, private cloud, or fully air-gapped, and IREX has no data plane on the other side of it.
IREX neither owns nor accesses customer video, events, logs, alarms, watchlists or floor plans. There is no IREX-side data plane for customer media. That is an architectural fact, not a contractual promise.
On-premises, private cloud, or fully air-gapped. There is no internet dependency at runtime, and IREX-hosted SaaS is an opt-in convenience, never a requirement.
The infrastructure stack is entirely open source, with no third-party proprietary code: Kubernetes, Ceph, Cassandra, Kafka, PostgreSQL, Elasticsearch. IREX-authored components are available for source review or escrow under license.
Every high-risk AI action is bound to a mandatory Case ID and written to an append-only, tamper-evident log, exported daily, with read access for supervisors and independent oversight bodies by design.
See Security & Compliance → Read the Data Sovereignty Guide ↗
What We Build
IREX is camera-agnostic, covering ONVIF Profile S and the major brands with no rip-and-replace, and ships the world's most powerful video analytics platform under one contract, one SLA and one upgrade path. That means the world's first VLM analytics for public safety (StreamVLM™, prompt-defined detectors, beta on selected instances), the broadest set of specialized analytics modules on one platform, spanning faces, vehicles and traffic, weapons, perimeter, rail and transit, crowds, fire and smoke, camera tamper, and access biometrics, and a petabyte-scale instant video search engine.
Instant search across petabyte-scale video. Results come back in under a second, from billions of face, figure and vehicle objects, by map, filter or plain language. Partners report 100x+ reductions in investigation time and 10x+ in response time, with 1,000+ cameras per operator.
Learn More →Vision-language analytics you define by writing a sentence. No model training, no dataset collection. Multiple prompt-defined detectors per camera, for the long tail no vendor ships out of the box: a person lying down, flooding, illegal dumping, graffiti, uncleared snow. Shipping in beta on selected instances.
Learn More →The agentic layer. Describe an investigation in plain language and IREX agents interpret the intent, cross-reference watchlists and run the multi-step search. That works from the platform or from the Sover secure messenger, inside the same permissions, Case ID and audit trail as any other search. Shipping in beta on selected instances.
Learn More →Before any privacy-sensitive AI action, the operator records the lawful grounds for it: a case file, court order or missing-person report. A world first in video analytics, and a platform-level mandate rather than a setting an administrator can switch off.
Learn More →One instance is proven at 100,000+ cameras, and the architecture scales linearly across cameras, users, retention and active AI modules alike, up to an architectural ceiling of 1,000,000 cameras and 100,000 users. Every component is hot-swappable; there is no single point of failure.
Learn More →End-to-end encrypted messaging, voice and video for jurisdictions that need to share evidence and coordinate across agency lines. It is user-hosted, and the secure front end through which alerts arrive and Ask IREX answers.
Learn More →Where We Operate
Headquartered in Murrieta, California, with a distributed team of about 100 people: executive leadership in the United States, Andorra and the UAE, and regional delivery teams across LATAM and Europe. The eight countries below are production deployments; we count nothing else.
Southern California law enforcement: traffic, weapons and watchlists; selected for smart-city deployments across San Diego County.
London Underground: railway safety, crowding and unattended items; perimeter security on UK critical infrastructure.
Dublin Airport: watchlist suspect recognition, unattended luggage, weapon detection and crowd management.
National Police and the Lima / Miraflores multi-city command centers, with a large-scale national rollout in partnership with Visium.
Retail chains: watchlist-based loss prevention.
Every Premier League national stadium: banned-fan watchlists and crowd management.
The first nationwide multi-camera traffic-enforcement deployment, including Ulaanbaatar.
A national smart-city program running the full IREX module catalog.
The Team
IREX is a globally distributed team of 100+ people, more than 50 of them software engineers and data scientists, including a ~30-person machine-learning group behind the analytics modules and StreamVLM. Below are the people customers and partners work with directly.
The two co-founders lead the company and the platform.
Co-Founder & Chief Executive Officer
Global strategist across public-safety technology, intelligence/surveillance/reconnaissance programs, AI-ML IoT ventures and critical infrastructure. Executive sponsor on every IREX engagement.
Co-Founder, President & Chief Technology Officer
Computer scientist and serial founder, with a PhD in machine learning and cryptography, 50+ peer-reviewed publications and 10+ granted patents in computer vision. Leads product and engineering direction.
Independent advisors, not full-time IREX personnel.
Advisor, Ethics and Compliance
Former Executive Director of INCITS, heading the US secretariat for international IT standards under ISO/IEC JTC 1. Founding Executive Director of the Technology Integrity Council. Advises on biometrics, AI governance and public-safety data handling.
Advisor, Partnerships
Global business network and technology advisory boards, with an NCS4 spectator-safety background. Advises on constitutionally-aligned surveillance AI for governments.
Advisor, Investments
35+ years in venture capital across AI, SaaS, privacy and security companies. General Partner at Hidalgo VC.
The people who build, scope, deploy and support every IREX program.
VP of Engineering
Leads the IREX engineering organization: 50+ engineers and data scientists across back-end, front-end/UX, DevOps/NOC, QA and support.
Director of Operations & Deployment
Solutions architect and IT leader, previously in enterprise physical security for a global financial institution and Assistant Director of IT at a Tier-1 public-safety agency. Owns end-to-end solution design.
Business Development
Master's degree in information technologies, quadrilingual, 8+ years in AI engineering. Now the working-level bridge between IREX engineering and customer operations.
Head of Public Relations
Runs communications end to end, leading a PR team of 10+: traditional and online media relations, social media, and video production.
Head of Sales
Programmer and business analyst turned sales lead, with deep experience in global software sales and implementation across regions.
Public Sector Partnerships
National-scale safety and digital-sovereignty programs: strategy, delivery and market entry across B2B and B2G.
Business Development, LATAM
Smart-cities advisor and AI evangelist, 25+ years as a software architect. Drives IREX growth across Latin America.
Technical Lead, LATAM
8+ years leading AI video analytics, smart-city and public-security technology projects across the LATAM region.
Marketing
Demand generation and account-based marketing for public-safety technology.
Social Impact
A missing child is not a procurement line item. IREX partners with non-profit organizations that search for missing children and adults, the National Child Protection Task Force among them, so that the technology built for citywide public safety is available to the people doing that work, and we support schools working to curb drug distribution.
Those partnerships are also the sharpest test of the ethics framework: the capability that finds a trafficking victim is the same capability that could be misused against an ordinary person. The Case ID mandate, the pre-registration constraint and the audit trail exist so that the first is possible and the second is not.
Ecosystem
IREX works with two kinds of partner, and the difference is worth stating plainly. Technology partners extend what the platform can do and how it is delivered. Mission partners put it to work on cases that were never a commercial opportunity.
Certified suppliers that fill capability gaps around the platform: hardware, hosting, integration and analytics. Each has passed the IREX technical certification process, and the ecosystem is certified rather than open. Alongside it runs a channel, distribution and system-integrator program for delivery in-region, under which IREX still contracts with the end user wherever that is legally possible.
Non-profit organizations that use the platform to find missing children and adults and to disrupt human trafficking. There is no commercial relationship in either direction: IREX provides the technology and the datasets, they provide the casework, and the arrangement exists because that work does not pay for itself.
Keep Reading
Tell us about your camera environment, deployment constraints and compliance requirements, and we'll arrange a walkthrough on your own footage. Not a canned demo.