The SurveillX platform
From field signal to operational action.
Capture the scene, add environmental and system context, preserve the best evidence, and put the whole field into the hands of operators, software, and governed AI agents.
Capability map
One platform. Twelve connected layers.
Core capabilities ship broadly. Hardware and third-party paths marked “deployment-specific” require the relevant device, credentials, or configuration.
Continuous capture at the edge.
SurveillX records directly where the cameras are. The stream pipeline records H.264, H.265, and AV1 camera video locally as standards-based fragmented MP4 (CMAF), monitors continuity, and recovers from the unstable camera and storage conditions that make remote deployments difficult.
Responsive live view across changing links.
The default live path delivers each camera as its own sub-second stream over Media over QUIC (WebTransport)—a next-generation transport built for lossy cellular links that scales to many viewers without multiplying the site uplink. WebRTC composited sessions, cloud fan-out, adaptive HLS, and a snapshot grid cover every other condition.
Detect locally. Analyze deliberately.
Edge inference handles the continuous work; higher-level vision analysis is invoked when a human or scheduled workflow needs an explanation or comparison. That keeps the system useful offline and applies model cost where it creates an answer.
Turn activity into searchable, cited answers.
Panoptic builds a customer-scoped intelligence record from activity windows, events, and selected images. On Cognia-configured sites, learned baselines, anomaly signals, novelty selection, and coverage semantics add richer evidence without replacing the source archive.
Move from alert to proof.
Every investigation begins with time. SurveillX indexes the recording archive so teams can jump to a day, event, range, or frame, then protect and export exactly what the incident requires.
Know what surrounded the frame.
Power, weather, light, location, sensor state, and connectivity often explain the event as much as the pixels do. SurveillX brings these signals onto the same edge node and exposes them through one operational surface.
Operate cameras and the physical site.
SurveillX extends beyond passive viewing with deep camera onboarding, PTZ automation, live audio, and optional digital I/O. These controls let one edge platform coordinate what the site sees, says, and triggers.
Compress a changing site into a useful story.
Timelapse is a first-class archive, not an afterthought. Capture policy can follow time and available light, while frames remain searchable, protectable, shareable, and available for configured AI comparison.
Operate the infrastructure remotely.
Distributed field systems need more than a camera list. SurveillX exposes the state of the application, recorder, storage, GPU, networking, power, tunnels, and deployments so operations teams can diagnose before they dispatch.
Give AI tools, context, and guardrails.
Model Context Protocol turns the fleet API into a language-compatible operational surface. An assistant can gather evidence across systems, summarize findings, and propose or execute permitted actions without hiding the underlying data.
Fit into the systems around the site.
SurveillX is designed as a field intelligence layer—not a closed camera island. Its REST API, signed events, camera protocols, alarm transports, cloud backends, and construction workflows connect evidence to existing operations.
Put the whole site in front of the right people.
Operators work in the browser-based SurveillX console; account administrators manage sites, people, and alerting in the customer portal. Both sit on the same access controls as the API.
Architecture
Keep the hot path local. Connect everything else deliberately.
Scene + state
- RTSP / ONVIF cameras
- CSI + USB cameras
- GPS + weather
- LoRaWAN + digital I/O
- Power + cellular
Record · learn · index · act
- GStreamer / DeepStream / TensorRT
- Multi-database local timeline
- Optional activity learning + selection
- Health, retention, and schedulers
- REST, WebSocket, and MCP tools
People + systems
- Prism operator experience
- Panoptic search + reports
- Azure / S3 backup
- Webhooks + alarm stations
- AI assistants + fleet agents
Common questions
Platform, plainly.
No. Core recording, retention, local live view, native object detection, health monitoring, and many site actions run on the edge. Connected services add remote access, backup, fleet aggregation, external notifications, and model-backed analysis.
Yes. SurveillX supports standard RTSP and ONVIF IP cameras, including deeper integrations for several major vendors. CSI and USB cameras are also supported on compatible edge hardware. Exact camera and PTZ behavior is validated during deployment design.
No. Weather stations, LoRaWAN devices, Victron systems, Peplink routers, and GPIO require the corresponding hardware and configuration. SurveillX exposes a common operational model, while each physical deployment is validated against its equipment.
No. Optional SHA-256 hashes and protection flags support file-integrity and retention workflows, but they are not presented as a certified immutable chain-of-custody system. Formal evidentiary requirements should be scoped separately.
Design the deployment
Start with the site, not a SKU sheet.
We’ll map the field conditions, evidence goals, operator workflows, cameras, sensors, and integrations into a deployment that can be tested.