Built into the platform
TetherX AI
Traditional motion detection tells you something moved. TetherX AI tells you what moved. Video AI identifies people, vehicles and animals. Audio AI classifies 63 security-relevant sounds. Describe the scene. Find the evidence using TetherBox AI where available, or the analytics and events already coming from supported cameras and connected systems. Bring alarm, access, sensor and relay context together with the footage. Count directional traffic and configure responses on compatible hardware.
Describe the scene. Find the evidence. Use available native AI, existing camera analytics and supported connected-system events across your permitted sites.
On-Device Detection
An AI-capable TetherBox handles video and audio detection locally, alongside recording, without a separate analytics server.
Describe the Scene
Find recorded evidence using available TetherBox AI or existing camera analytics, with supported security-system events adding context.
Two Senses
Video AI and audio AI run together on the same TetherBox - audio covers the corridors and spaces a camera cannot see.
Search Within Your Access
Search interpretation receives your request and authorised location names, not footage. Results respect your site, camera and footage-access permissions.
Choose the insights your site needs
Combine detection, appearance checks, counting and response on compatible TetherBox hardware.
Detect and track
Recognise people, bicycles, vehicles and animals. Follow movement across the scene and review the matching event.
Find by appearance
Use recorded clothing, hat, vehicle and animal colours to narrow an investigation across permitted cameras.
Check visible PPE
Enable hi-vis, hard-hat and helmet checks on suitable views. An unclear person remains unassessed; these checks do not certify safety compliance.
Count directional traffic
Compare entries and exits by class and hour, export CSV and open related events. Support configures the counting line.
Listen for activity
Classify recorded sound events such as glass breaking, screams and alarm tones. A suitable microphone or camera audio feed is required.
Configure a response
Set region-count rules for unattended service points or sustained gatherings, with selected notifications and compatible outputs.
Review the enabled modules, camera views and hardware requirements with your integrator. See setup and licensing questions.
We Taught TetherBoxes to See
Not just detect motion - actually understand what is in the frame. Video AI identifies people, vehicles and animals in real time and tracks their movement across the scene. Support configures a counting line on compatible AI hardware, with separate people, bicycle, vehicle and animal crossings, charts and CSV export.
Detection That Holds Up in the Real World
The images here are from a live site in Canada - night time, snow on the ground, well below zero. Two people detected at distance as tiny figures in a wide scene, at high confidence. Most camera analytics would not register them at all.
Each detection is tracked frame to frame, so you see the full trajectory: where someone came from, the path they took, and where they went.
People Counting at a Glance
Compare directional entries and exits over the last 30 days, or open a day hour by hour. Export CSV and open the events behind a figure. Support configures counting on compatible AI hardware; crossing totals are not unique visitors or verified occupancy.
Describe the scene. Find the evidence.
Tell TetherX what happened, where and when. Find the relevant footage using native TetherBox AI where available, or supported existing camera analytics and recorded events. Bring supported alarm, access-control, sensor and relay events into the investigation context.
- “Show me a person wearing a red hat last Tuesday.” Combines a person detection, hat colour and date.
- “Vehicles at all front doors at night.” Matches permitted front-door cameras and the overnight time window.
- “Line crossing events in June, mornings only.” Searches the whole month with a separate daily time window.
Person, Vehicle or Animal
Narrow results to the detection type you care about.
Clothing and Hat Colour
Search recorded top, bottom and hat colours, including a person wearing a red hat.
Vehicle and Animal Colour
Narrow recorded detections by colour across your permitted sites and cameras.
Moving or Stationary
Filter out parked vehicles, or focus on loitering.
Location and Time
Combine permitted sites and cameras with calendar ranges and daily time windows.
Recorded Analytics
Ask for supported detections such as line crossing, intrusion or parking.
Filter for a person in red, 6pm to midnight - results across every camera in seconds.
Know when people arrive and when an area needs attention.
TetherX Support configures counting and rules on compatible AI-capable TetherBox hardware. Cloud notifications and output commands need connectivity. Crossing totals are not unique visitors or a verified occupancy count.
People counting
Count directional entries and exits. Compare busy hours, export CSV and open the recorded events behind a total.
Unattended counter
Alert when a staff position stays empty while customers wait. Notify the manager or play a recorded message.
People gathering
Notify the team when a people threshold persists in a drawn area. Set the duration and site-local active hours.
Sensors and relays
Put leak or contact alerts beside footage. Switch compatible lights or sirens, or pulse a control input with suitable wiring.
Rules check people detected in a region over time. Counting totals crossings over time. Choose push/email, recorded audio or compatible output actions, then review the activation event and action results.
We Taught TetherBoxes to Hear
Video analytics have a blind spot - they only work where the camera can see. Audio AI classifies 63 distinct sounds across 7 categories, running continuously as a second sensory layer.
Detecting What Cameras Miss
Glass breaking behind a wall, a smoke alarm in an adjacent room, someone screaming off-screen - none of these trigger video analytics. A site with 16 cameras still has corridors, stairwells and spaces with no visual coverage.
Audio events appear with distinctive purple icons on event thumbnails, instantly distinguishable from video detections when an operator is scanning dozens of events. Loud sounds above 99 dB SPL are automatically flagged regardless of classification.
Filter by category or individual sound class, and route notifications by type - glass break alerts to the monitoring station, dog barking to the site manager.
Seven Categories, 63 Sound Classes
Emergency
Screams, cries for help, shouting
Alarms
Smoke detectors, fire alarms, security sirens
Intrusion
Glass breaking, door forced, wood splintering
Vehicles
Engine start, horn, tyre screech, reversing beeper
Voice
Speech, arguing, crowd noise
Animals
Dog barking, bird of prey
Weather
Thunder, heavy rain, high wind
individual sound classes across every category
Two Sensors Create Context Neither Has Alone
A person detected on camera at 3am is interesting. A person detected at 3am with glass breaking on audio is an incident. Video AI and audio AI run on the same suitably equipped TetherBox. A compatible video feed and microphone or camera audio feed provide the two inputs for the operator to review together.
Video and Audio Detection Stay On-Site
Video and audio detection run on an AI-capable TetherBox alongside recording, without uploading footage for those detections. Natural language search is a cloud service that interprets your request and authorised location catalogue; it needs an internet connection.
No Extra Server
Detection runs alongside recording on a suitably equipped TetherBox. Your integrator checks accelerator capacity and the modules you need.
Detect Without Uploading Video
Detection happens where the cameras are. Search requests use the connection to interpret your words and retrieve matching events.
Data Sovereignty
Video and audio detection stay on-site. Review cloud search, remote viewing and optional recording requirements alongside your data policy.
Cloud Backup Still Works
Optional cloud recording continues exactly as before, per camera.
AI software. Your choice of compatible hardware.
Counting and rules need compatible AI-capable TetherBox hardware and software, a suitable camera view and Support configuration. Counting access follows your site and camera permissions. Rules need configured outputs and recipients; cloud notifications and commands require connectivity. Ask Support to check the cameras, regions and response you want.
Use suitable NVIDIA GPU or Jetson hardware, or a Raspberry Pi with a compatible Hailo AI HAT. TetherX AI runs as software on the TetherBox, so you can choose supported cameras and compute for the job. Each installation needs a check of the accelerator, enabled modules, camera workload and software compatibility.
| Hardware choice | What your integrator checks |
|---|---|
| NVIDIA GPU or Jetson | Compatible CUDA software, platform runtime, model support and sustained camera workload. |
| Raspberry Pi with Hailo AI HAT | The exact Hailo chip, compiled models, enabled modules and camera workload. |
| Existing site hardware | Whether a supported accelerator and sufficient recording capacity make it suitable for the required AI. |
Choose the platform with your integrator, then confirm the setup in the AI hardware requirements and TetherBox deployment options.
Full Details & Hardware RequirementsTetherX AI questions
Yes. Support configures directional counting lines and people-region rules on compatible AI hardware. Compare daily or hourly crossings, export CSV and open related events. For rules, choose a region, people threshold, duration and active hours, then selected notifications or compatible outputs. Cloud notifications and commands need connectivity.
Crossing totals are not unique visitors or a verified occupancy count. Region rules check the people detected in an area; missing observations do not establish that it is empty.
Video and audio analytics run alongside recording on a suitably equipped TetherBox. Your integrator checks compatible accelerator hardware, the selected modules and camera workload. Existing site hardware may be suitable; an upgrade can be required.
Yes. TetherX AI runs as software on compatible NVIDIA GPU or Jetson hardware, or Raspberry Pi hardware with a compatible Hailo AI HAT. Your integrator checks the platform runtime, exact accelerator, enabled modules and camera workload before installation. Read the hardware choices and setup requirements.
Yes. Describe the scene in Find footage, such as a person wearing a red hat last Tuesday. TetherX uses the supported analytics and recorded events available across your permitted sites, whether supplied by native TetherBox AI or existing cameras. Supported alarm, access, sensor and relay events provide wider incident context.
The preview and editable dates let you check the interpretation when needed. Results depend on the events and attributes actually recorded; a new native-AI deployment is not required for every search.
Video and audio detection run locally on suitable TetherBox hardware without uploading footage for classification. Natural language search sends your request and authorised location catalogue to a cloud interpreter, not the video. Remote viewing retrieves the clips you open, and optional cloud backup uploads footage for cameras you select.
The TetherX platform subscription is from £100/site/year. TetherBox AI is included in the platform licence; additional modules and data retention may cost extra. Your local integrator scopes suitable hardware and any optional cloud recording, specialist analytics or professional monitoring services for the site. Request a quote for your camera estate and the workflows you need.
The AI runs on the TetherBox, not in the camera. Supported cameras can feed TetherX AI without an AI-specific camera SKU. TetherX supports 1,000+ integrations; your integrator checks the model, stream quality, camera view and suitable accelerator hardware for the selected AI modules.
Native video and audio detection run on suitable TetherBox hardware. TetherX also uses supported existing camera analytics and integrated events, so native AI is one input rather than a prerequisite for every investigation. Cloud interpretation turns your scene description into a supported evidence search. Read our cloud / edge / camera analytics breakdown for the processing choices.
We reply within 1 business day from a real engineer, not an SDR. Quote requests are routed to a local TetherX-certified integrator who handles the install and ongoing support.
Compare TetherX, honestly, to the alternatives.
Side-by-side pages against every cloud VMS, on-prem incumbent and manufacturer-bundled VMS integrators ask about - built from primary-source pricing, integrator forum quotes and current product documentation.
See all comparisonsDrowning in motion alerts that turn out to be a fox or a flag?
Describe the scene and find the evidence across your sites, using the analytics and connected systems already available.
Request a quote and a TetherX integrator scopes AI for your sites.
Try TetherX AI on real sites
TetherBox AI is included in the platform licence; additional modules and data retention may cost extra. A local TetherX integrator will check suitable hardware, enabled modules and counting terms, then scope any optional cloud recording, specialist analytics or professional monitoring for your sites.
We reply by email - usually within one business day.
Match me with a partner
Before you go - see TetherX AI on a real site near you.
Drop your work email and we will match you to an accredited local integrator running AI today.
No sales call; your email stays with us.
