Codecs stop being a blocker
If your NVR can display the feed over HDMI, camera-side H.264, H.265, and other NVR-supported formats do not need to be reconfigured for AI.
Analyze the final image →CODEC-INDEPENDENT VIDEO AI
Keep your cameras.
Let AI understand
what they see.
H.264, H.265, or any other codec your NVR already displays—it does not change our analysis. We read the final HDMI output and run detection on a managed GPU Edge, with no fleet replacement, codec migration, or inbound firewall port.
✓ Managed GPU Edge✓ Keep your current NVR✓ Traceable evidence and reasons
WHY IT WORKS
Keep the video system.
Start proving results.
This is not another camera fleet. It is a searchable, explainable AI layer added after the HDMI output of the NVR you already use.
Keep existing hardware
A compatible KVM reads the current NVR multiview. There is no camera-by-camera replacement and no need to expose every stream to our platform.
Reduce retrofit cost →Notify on outcomes
Combine people, vehicles, regions, crossings, dwell time, and scene state into explainable rules. Emails include the image, time, and reason.
Prove it with real results →Search the scene
Search retained representative images using object labels and semantic matching, such as suitcases or a person in a black shirt. Availability follows your account retention policy.
Account-based retention →THREE STEPS
Connect the picture. Then describe
the outcome that matters.
CONNECT
Connect the existing NVR
Our managed Edge reaches a compatible KVM through an authorized Tailscale network or reachable address and maintains a native WebRTC session.
DEFINE
Name views and conditions
Split the HDMI multiview, choose objects, and add regions, direction lines, or scene references when needed.
ACT
Search, verify, act
Receive an annotated image and reason when an event occurs, or search representative images later.
PRIVACY BY DESIGN
Continuous recordings stay on your NVR.
The managed Edge handles only the evidence it needs.
01Minimized retentionSearch retained representative images using object labels and semantic matching, such as suitcases or a person in a black shirt. Availability follows your account retention policy.
02Explicit viewing scopeUsers can only see the devices and views assigned by their administrator.
03Isolated managed EdgeTailnet identities, credentials, container networks, and evidence storage remain isolated between customers.
ANNUAL PRICING
Simple annual pricing
Free
US$0
1 free local channel · no semantic search
Local analysis
US$3 per camera / month
US$36 per camera / year · billed annually
Semantic search included
Only for processors owned by this account.
Hosted analysis
US$8 per camera / month
US$96 per camera / year · billed annually
Semantic search included
For processors shared by other accounts, or your own processors.
Unlimited rule events · 30-day cloud retention
USD, billed annually and renewed automatically until canceled. Cancel before renewal to stop the next annual charge; access continues through the paid period. Any applicable tax is shown at checkout.
Camera capacity
Buy, redeem and manage channels in one place.
Rule events and historical keyframes have no hourly count or aggregate image-byte quota. Keyframe uploads require no paid plan. Image size, API request rates, sampling, local cache and retention remain bounded; semantic search still depends on your plan. This is not unlimited video storage or guaranteed delivery.
SUPPORT GUIDES
Guides & help
Set up detection, understand rule boundaries, find past evidence, and connect notifications. Every guide is public and has its own shareable link.
Browse all guidesRules & detection
RULES
Create rules and receive alerts
Choose cameras, conditions, per-camera scopes and recipients, then understand events and upload status.
Read the guide RULES & REGIONSRule conditions, regions & exclusions
Choose the right trigger and camera scope, and understand when masks and foreground priority take effect.
Read the guide HARDHAT / PPEDetect missing safety helmets
Configure person detection, a suitable area and multi-frame PPE checks before enabling helmet alerts.
Read the guide YOLO OBJECT DETECTIONYOLO objects & confidence
Select what to detect, read confidence correctly, and distinguish object labels from behavior.
Read the guideSearch & history
Message delivery
Deployment & privacy
START WITH WHAT YOU HAVE
Create free account →Put the first view to work.
Create an administrator account and we will arrange managed Edge onboarding for your NVR.