A documented comparison for security teams evaluating identity-led video analytics. Oosto is built around recognising specific people, which is a different premise from the rest of this market and carries a different review process. The alternatives here answer the same operational questions without it.
Oosto sells facial recognition technology used by security teams for persons of interest monitoring, video analytics and access control, available via cloud, on premises or SDK, with edge processing and image data converted to vectors. It also publishes behaviour detection for sitting, standing, fighting, running and slip and falls, plus touchless access control that leverages existing cameras. Teams shop for an alternative when identity recognition is not permitted or not wanted, when the questions are operational rather than who-is-that, or when the platform has to act on site. The five below answer those without an identity database.
Both routes put AI on cameras a site already has. They differ on whether the answer is an identity or an event, and that difference decides which approvals the project needs.
Every cell is something the vendor publishes, or an explicit not publicly specified.
Swipe the table sideways to see every column.
Oosto's row was read off oosto.com and its products page on August 12, 2026, including the deployment options, the vector conversion statement and the behaviour detection list. The other five rows come from each vendor's public documentation reviewed in July and August 2026 on the same five columns. Gaps are marked as not publicly specified rather than guessed.
Start with what Oosto is, because it is not the same product as the rest of this page. Its published premise is recognising specific people: real-time watchlist alerting, person of interest identification in crowded scenes and difficult lighting, and touchless access control that opens a guarded entry for an authorised person using existing cameras. It publishes deployment via cloud, on premises or SDK, edge processing, and the technical detail that images are converted to vectors including at the edge. It also maintains a published ethics and accuracy position, which is the right instinct for this category.
The first real reason teams look elsewhere is governance, and it usually arrives before the technical evaluation. An identity-recognition deployment brings a legal basis, a retention policy, a works council or employee consultation, a data protection assessment and in several jurisdictions a hard limit on what is permitted at all. None of that is a criticism of the product. It is a project that runs in parallel with the software one, and for many teams the honest conclusion is that they wanted to know what happened rather than who it was.
The second is operational scope. Watchlist alerting answers who, and the behaviour set adds fighting, running and slip and falls. It does not answer whether the dock was blocked for forty minutes, whether the crew wore harnesses, how long the queue ran at shift change, or what to do at 02:00 with nobody watching, since no on-site deterrence behaviour is published. Those are the questions the alternatives here are built around, and none of them needs to identify a person to answer them.
Ask what question you are actually funding. Who was that is a different purchase, and a different approval process, from what happened and what should happen next.
What each one is, where it is strong, and what to check before you commit.
Spot AI is a software-led video AI platform that turns the IP cameras a business already owns into AI coworkers. An on-site Intelligent Video Recorder keeps full-resolution video in the building and sends only event metadata to the cloud, so search and multi-site management stay cloud-based while footage does not leave the facility.
Eagle Eye Networks is a cloud video management system that connects to virtually any ONVIF-conformant camera and digitises analog feeds through analog-ready Bridges and CMVRs, then layers cloud analytics on top of the estate. Its documentation names ONVIF Profile S and dual codec streaming as the integration path.
Cobalt AI sells Cobalt Monitoring Intelligence, an automation layer that connects to the cameras, access control systems and edge devices a site already runs, then reasons across them to surface root cause and drive response workflows. It supports edge, cloud or hybrid deployment and is built on over ten years of in-house security operations experience.
Verkada is a cloud-managed platform that sells its own cameras, access control, sensors, alarms and intercoms under one console, and publishes per-device MSRPs, which is unusual in this market. Third-party cameras come in through Command Connector, which Verkada states is an ONVIF Profile S conformant client and can also ingest RTSP feeds.
Rhombus sells cloud-managed cameras, sensors, access control and alarm monitoring as one system, described by the company as a cloud-edge system that operates offline. Third-party cameras come in through the Relay Connector N100, and the company publishes SOC 2, NDAA, GDPR and TAA compliance badges.
Six questions that separate these platforms faster than any feature list.
Write the requirement as a sentence. If it reads recognise this named person when they return, that is identity recognition and this list narrows to Oosto. If it reads tell me when someone is in the yard, when a forklift comes close to a pedestrian, or when the dock is blocked, the event is the requirement and every alternative here answers it without an identity database.
Settle this before the demo, not after. Identity recognition typically needs a documented legal basis, a retention policy, a data protection assessment and employee consultation, and in some jurisdictions it is restricted outright. Oosto publishes an ethics and accuracy position, which helps the conversation, but the obligation stays with the deployer.
Oosto states its access control product leverages existing cameras, without publishing a protocol or conformance profile. Eagle Eye publishes more than 7,500 models plus analog through Bridges. Spot AI takes any ONVIF or RTSP camera and legacy analog through the Intelligent Video Recorder. Recognition workloads are also sensitive to angle and lighting, so pilot on your worst-placed camera rather than your best.
Oosto publishes cloud, on premises or SDK, edge processing, and image data converted to vectors including at the edge, which is a meaningful answer to a data protection reviewer. Ask every vendor the same three questions: where the video sits, where the inference runs, and what derived data is retained and for how long.
Ask what the match triggers at 02:00 when no operator is at the console. Eagle Eye documents sirens and talk-down alerts. Verkada gates automated audio to the BZ11 or an Intercom. Spot AI runs talk down, strobes and horns through standard speakers already mounted. Oosto, Cobalt AI and Rhombus publish no on-site deterrence behaviour, so a person closes the loop.
If the list includes PPE compliance, forklift near-misses, hazard-zone crowding or dock delays, ask which of those ship pre-trained. Oosto's published set is identity and behaviour shaped, and none of the vendors here publishes a full operations and safety agent set except Spot AI, so that is a question the others route to a partner product.
Spot AI fits when the questions are what happened, what it means and what should happen next, answered from cameras already mounted and without an identity database. It runs on any ONVIF or RTSP camera plus legacy analog through the Intelligent Video Recorder, keeps full-resolution video in the building, ships 15+ pre-trained agents across security, safety and operations, and acts through standard speakers already on site.
Oosto stays the right call for an organisation whose requirement genuinely is recognising named individuals in real time, that has the legal basis and the governance to run it, and that values published deployment choice across cloud, on premises and SDK. On that specific job the alternatives here are not substitutes, and this page does not pretend they are.
A camera-agnostic platform lets you test the AI on your own cameras before any hardware decision, which is the cheapest way to de-risk the choice.
Watch the AI coworkers work on your live feeds, not a demo reel.
Customer-reported outcomes from named Spot AI customers.
Storage Asset Management was alerted to a break-in around one in the morning, and law enforcement arrived while the incident was still in progress.
Don Franklin Family of Dealerships had video of an incident on responding officers' phones within four minutes of the alarm, and recorded five recoveries within the hour.
Cambridge City cut the time it takes to find footage from two hours to thirty seconds.
"The Spot AI notification went out around one in the morning, and police were able to arrive on-site while the incident was still in progress."
The five covered here are Spot AI, Eagle Eye Networks, Cobalt AI, Verkada and Rhombus. Spot AI fits teams whose questions are events and actions rather than identities. Eagle Eye Networks fits mixed and partly analog estates with named analytics. Cobalt AI fits automated triage across existing systems. Verkada fits a site specified from scratch. Rhombus fits mid-market teams wanting one vendor.
Not as their premise, and that is the point of the comparison. Eagle Eye names Face Match among its analytics. The others on this page are built around events, behaviour and operations rather than identifying named individuals. If identity recognition is a hard requirement, Oosto is the platform on this page designed for it.
Its published capability is identity led: real-time watchlist alerting and person of interest recognition, described as working in crowded environments and difficult lighting, plus touchless access control using existing cameras. It also publishes behaviour detection covering sitting, standing, fighting, running and slip and falls. No on-site deterrence behaviour appears on the public pages.
It publishes cloud, on-premises or SDK deployment, states it brings security to the edge, and describes data encryption that converts images to vectors including at the edge. Those are the answers a data protection reviewer will ask for, so get them confirmed in writing for your configuration along with retention periods for the derived data.
No. No price, plan tier or contract term appears on its site, which is the norm across this category. Of the platforms compared here only Verkada publishes per-device MSRPs. For an identity deployment the software quote is also not the whole cost, because the governance work runs alongside it.