DICE Corp. has announced a new edge artificial intelligence platform designed to run AI processing within a customer’s network and report results directly into Ceila.ai as a native component.

The platform was previewed with corporate and national account clients at GSX 2026 in Atlanta and is currently undergoing trials with monitoring centers and enterprise accounts, including interactive virtual guarding provider Clear Watch. General availability is planned for the first quarter of 2027.

The new platform provides an alternative to video monitoring deployments that rely on third-party analytics providers. In those deployments, motion events are sent to an outside provider for evaluation before the results are forwarded to the monitoring platform.

With the DICE edge platform, processing instead takes place on site using DICE hardware. Results are delivered directly into the workflow handling the alarm signal, video, audio and dispatch. Video clips do not leave the customer’s premises for evaluation.

DICE said the local processing approach also simplifies network configuration because the appliance operates inside the customer’s network and communicates outbound to DICE. Cameras are accessed across the local area network rather than through inbound access.

The platform is designed to carry DICE’s full detection library on premises. Capabilities include person and vehicle classification, behavioral and object interaction detection, license plate recognition, people and vehicle counting, privacy masking, thermal imaging processing and safety and compliance review. The latter includes detection related to personal protective equipment use and lifting technique.

DICE also develops detection models for individual customers, including custom projects for Fortune 500 companies, national retailers and municipal operations.

The company said development of the platform has been driven by national retail chains and multi-site enterprise accounts where event volume and bandwidth costs make on-premise filtering important. DICE also said it has seen an increase during the past year in monitoring centers and guarding operations migrating to its platform and has expanded installation and onboarding capacity.

Clear Watch currently runs its AI video detection on the DICE platform.

“In virtual guarding, our reputation is built on catching what matters and staying quiet about what doesn't. Three object classes never gave us the resolution to do that reliably,” said Glen Carter, president of Clear Watch. “Running DICE's full detection library on the edge means the call gets made on site, in real time — we're not dependent on a third party's turnaround to protect our clients.”

DICE remains camera-agnostic, with the platform built to established industry protocols including ONVIF. Devices outside those standards can connect directly through RTSP, with events delivered into the monitoring center through SMTP. DICE said it will also continue supporting third-party analytics providers selected by customers.

Ceila.ai extends DICE’s digital operator and co-pilot architecture to virtual guarding providers and enterprise global security operations center teams. DICE manufactures its own PBX and softswitch, allowing Ceila.ai to evaluate live phone audio and live video within a single decision path.

Dispatches are delivered through the DICE message broker, which is connected to more than 700 public safety answering points and sends alarm dispatches directly into their computer-aided dispatch systems.

“We built Ceila.ai with AI at the core of the event management platform rather than as a service we forward events to,” said Cliff Dice, president, CEO and founder of DICE. “If a customer wants to bring their own provider, we integrate it and we do that every day. But they should also be able to get it from the same company that handles their alarm, their video, their phone call, and their dispatch, with one throat to choke when something misses.”

The DICE platform runs on Linux using a runtime engine and virtualization layer developed by the company inside UL-listed hosting centers that DICE owns and operates. The architecture is designed to allow processing capacity to expand and contract automatically based on load.

DICE said it provides computing capacity within its own centers and incorporates that cost into the platform rather than relying on analytics pricing established by a third-party provider.

The company has invested millions of dollars in infrastructure supporting the program, including its own GPU computing environment within facilities it owns and operates, according to the announcement. DICE has also filed more than 20 patents related to Ceila.ai, adding to a patent portfolio that includes its Video Arm/Disarm and Gatekeeper technologies.