Loading…
Loading…
Cloud, on-premise, edge, or hybrid. Visylix adapts to your infrastructure. Run AI video analytics wherever your cameras are, with container-native orchestration and hardware-accelerated processing at every layer.
Choose the deployment model that matches your security posture, latency requirements, and operational preferences.
Run Visylix in your own AWS, Azure, or GCP account. You keep control of the infrastructure and the data; we ship the image and the deployment manifests.
Deploy Visylix on your own hardware for complete data sovereignty. Ideal for air-gapped environments and strict compliance requirements.
Run AI inference directly on edge devices near cameras for fast local response with no WAN round trip, and reduced bandwidth consumption.
Combine edge sites with a central deployment. Run time-critical inference locally, federate every site into one view, and tier long-term footage to your own S3-compatible storage.
Run Visylix on a fully isolated network. Offline licence activation, local AI and a local copilot mean your licensed analytics and Radha run with zero external calls.
Every deployment model shares the same robust, secure, and performance-optimized infrastructure layer.
Fully containerized across 8 core services, with Kubernetes manifests for orchestration. One-command installation with an auto-tuning installer. Scale individual services independently based on workload demands.
Purpose-built for GPU acceleration from edge compute devices to data center GPUs. GPU-accelerated AI inference in every deployment mode, plus on-the-fly H.265 to H.264 transcoding for browser live view.
Mutual TLS between federated sites, encrypted credentials at rest, and role-based access control with audit logging for every operation.
Rolling updates protected by Kubernetes pod disruption budgets. Update AI models without interrupting live streams.
Federate sites across regions for data residency and edge proximity, with a primary-standby pair for recovery in under 60 seconds.
Architecture designed for SOC 2, GDPR, HIPAA, and ITAR compliance with configurable data handling and retention policies.
Our solution architects will help you choose and configure the optimal deployment model for your infrastructure, scale, and compliance needs.
Visylix supports five deployment modes: cloud, on-premise, edge, hybrid, and fully air-gapped. Installation is a one-command container deployment with an auto-tuning installer that profiles your hardware, orchestrating 8 core services, with Kubernetes manifests available. It runs on bare metal, virtual machines, GPU clusters and edge compute devices.
Yes. Visylix supports air-gapped and secure government-grade deployments with on-premise infrastructure, no external dependencies, and full offline operation capability.
No. Visylix is software delivered as a container image. You run it on your own servers, virtual machines or your own cloud account. We do not sell cameras, appliances or hosting, which is a deliberate difference from vendors that bundle proprietary hardware.
Yes, and this is the part worth checking with any vendor. All 22 AI analytics and the Radha copilot execute locally on your hardware. Many platforms describe themselves as on-premise while still calling out to a hosted service for inference or the assistant. Visylix does not, which is what makes genuine air-gapped operation possible.
Yes. The same container image runs in all five modes, so moving from a cloud pilot to on-premise, or adding edge sites to a central deployment, is a configuration and infrastructure change rather than a different product or a new licence.
In most cases. Visylix discovers ONVIF cameras automatically and connects to RTSP directly, so it is camera-agnostic rather than tied to specific hardware. Analogue cameras need an encoder to reach the network. NVR forensic access covers 15+ brands including OEM rebrands, so historical footage on existing recorders stays reachable during a migration.
It depends on camera count, resolution, codec and how many AI models you enable per camera, so we size it against your actual estate during the proof of concept rather than publishing a specification that would be wrong for most deployments. The auto-tuning installer profiles the hardware you provide and configures database pools, worker counts and buffer sizes to match.
No. Pricing is per plan, not per camera, per node or per deployment mode, so cloud, on-premise, edge, hybrid and air-gapped all carry the same licence terms. Plans start at $49 per month, with Indian pricing from Rs 4,599 per month, and Enterprise is quoted per deployment.