Visylix watches every camera with 22 AI analytics, decides what matters with Sudarshan rules, and acts in seconds: alerts, sirens, talkdown and escalation. Ask Radha anything in English, Hindi or Hinglish. It all runs on your own servers, and the VMS licence covers unlimited cameras.
One complete platform for live view, recording, AI analytics, automation and investigation. Every part of it, including the AI and the copilot, runs on your own servers.
Face recognition, number plates, weapons, fire and smoke, PPE, crowds, falls, intrusion, tailgating, queues, abandoned objects, tampering and audio events such as gunshots and screams. Each camera learns its own normal, and marking a false alarm tightens it automatically.
Ask in English, Hindi or Hinglish and Radha does the work. 111 tools to find footage by description, trace a plate across cameras, start a PTZ tour, enrol a face, write an automation or brief the next shift. Fully on premise, with confirmation on every high-impact action.
Live AI on the wall itself. 12 models draw overlays on the feed you are watching while the rest of the wall stays clean, and you arm a watch by describing it: "a red truck at the loading dock", "this plate", "anyone crossing this line".
Rules that combine conditions, time sequences and badge swipes, then act: alerts, siren and door-strike relays, talkdown through the camera speaker, webhooks, SIEM and escalation when a critical alarm goes unacknowledged.
Preview any rule against real history before you arm it. Every rule keeps a version history you can restore in one click, with snooze, hourly caps and maintenance windows. Every alert is scored, so noisy rules are retuned in one click.
Open recordings straight from Hikvision, Dahua, Uniview, I8H, Bosch and Panasonic recorders (NVRs), plus CP Plus, Honeywell, Sparsh and Godrej models. ONVIF, the open camera standard, brings in thousands of camera models.
Live video in under a second over WebRTC, down to under 200 ms. A native streaming engine built in-house handles 13+ protocols, including RTSP, RTMP, SRT, HLS, NDI and GB28181.
Export a self-contained player sealed with SHA-256 that opens offline, so anyone can check the footage is untouched. Faces, plates and custom zones are redacted on export. Recording keeps the camera's original quality without re-encoding.
On premise, edge, hybrid, your own cloud account or fully air-gapped. Large camera counts on standard servers, 500+ REST APIs and 55 interface languages, including 13 Indian languages.
Sudarshan is live AI on the video wall itself. It puts real-time intelligence on the feed your operator is watching, arms watches you describe in plain words, and turns detections into rules that decide and act on their own.
The feed you are watching gets rich live overlays across 12 models: people, vehicles, faces, plates and tripwires. Every other tile stays clean and lights up with a severity ring only when it needs you, so a large video wall never dissolves into noise.
Type what matters. "A red truck at the loading dock." "Anyone crossing this line." "Plate MH12AB1234." Sudarshan arms a live watch for faces, plates, plate patterns, objects in a zone, line crossings or anything you can describe in words.
Rules combine conditions, time sequences and access-control events, so "badge swipe with no matching face within 10 seconds" becomes a rule rather than a hope. Then Sudarshan acts: alerts, siren and door-strike relays, talkdown through the camera speaker, and escalation when nobody acknowledges.
Preview a rule against real history to see exactly what it would have caught before it goes live, then watch it on a rule health board. Version history with one-click restore, snooze, hourly caps and maintenance windows keep automation safe.
Every alert gets a dispatch score, and an effectiveness board shows which sources actually earn an operator response. Noisy sources are retuned in one click, within limits you set.
AI runs on exactly the cameras you choose, with detection zones per camera and per analytic. Everything is processed on your own servers, so no video or AI data ever leaves your building.
Watch for a red truck at the loading dock
Armed. I will flag any truck-class detection inside the Loading Dock zone and ring that tile the moment one appears. This stays active on the wall until you disarm it.
Alert me if anyone crosses the fence line, either direction
Watching Perimeter East. A line-crossing watch is armed on the fence tripwire for both directions, raising a warning-level signal with the direction of travel.
Tell me the instant you read plate MH12AB1234
Plate watch armed across 6 number plate cameras. I will show the camera, time and a snapshot the instant it reads, then hold the alert until someone acknowledges it.
Ask in English, Hindi or Hinglish and Radha operates your cameras, recordings, alerts and analytics for you. It runs entirely on your own servers and writes every action to the audit log. It is the only VMS copilot we know of that holds a conversation, takes action and runs fully on premise.
Ask in plain language and Radha gets it done. It starts recordings, pulls footage, traces plates across cameras, recalls PTZ presets, diagnoses cameras that have gone down, acts across the whole estate at once and works the alarm console, with 111 real tools.
Ask in English, Hindi or Hinglish. Radha understands follow-ups, remembers context, learns your site’s camera nicknames and takes spoken questions through fully local voice input. No commands to memorise, no menus to hunt through.
Every answer comes straight from your live cameras, recordings and system health, never from guesswork. For open questions Radha runs a multi-step investigation and returns its findings with a replayable evidence timeline. It can even look at a camera and describe the scene.
Radha stays loaded and ready on your own servers, so there is no waiting. It answers the moment you ask, every time.
A purpose-built language model that runs entirely on your own infrastructure. Your video and conversations never leave the building, and Radha works fully air-gapped.
Radha has no delete tool. Its 13 high-impact actions always ask for confirmation naming the exact target. Every tool respects the user’s permissions, and every action Radha takes is written to the audit log.
Why is the loading dock camera offline?
The camera is reachable on the network, but it is refusing the video connection, and the last 12 errors are all sign-in failures. That points to a changed camera password, not a network fault. I can restart the stream once the password is updated.
Kal raat Gate 2 pe koi laal gaadi aayi thi kya?
Haan. Ek red hatchback 11:42 PM pe Gate 2 se andar aayi aur 11:58 PM pe nikal gayi, plate MH12AB1234. Main dono clips aur poori plate journey khol doon?
Show me crowd alerts from today
Found 4 crowd detection alerts today: Cafeteria at 12:34 PM (78 people), Main Lobby at 9:02 AM (63 people), Auditorium at 2:15 PM (124 people). I can pull up the footage for any of these.
How many more cameras can this server take?
With the current headroom, roughly 40 more for recording only, or about 12 more if each runs one AI analytic. More analytics per camera would need a second GPU. I can break that down further if useful.
AI is only half of it. Visylix gives your security team everything it needs to run a site day to day, so nothing has to be bolted on from a second vendor.
PTZ cameras autotrack people and vehicles from live detections. One-click preset recall, patrol tours and a tour editor, and Radha can send any camera to a preset or start a tour by name.
Floor-plan maps with live camera status and alarm badges, plus a geographic map view on your own map tiles, so air-gapped sites keep their maps too.
A full operator console for acknowledging, escalating and closing incidents. Critical alarms that nobody acknowledges escalate automatically, and every alert carries a dispatch score.
Export an incident as a self-contained player sealed with SHA-256. It opens offline with nothing to install and anyone can check the footage is untouched. Built for police, prisons and courts.
Exact number plate journeys with travel times between cameras, plus person journeys that follow someone by appearance to narrow a search fast.
Sign in with Active Directory, LDAP, single sign-on or passkeys. 46 permissions across 4 roles, and a tamper-evident audit trail of every action.
ONVIF, the open camera standard, plus device packs for the top five camera brands. Add and configure cameras in bulk, use camera inputs and outputs, and take in camera events.
A device-health console shows failing cameras, storage runway and service state, so gaps are fixed before anyone needs the footage.
Signed webhooks, MQTT, SIEM feeds to Splunk and Datadog, PSIM links to Genetec Mission Control and Bosch BIS, and 500+ REST APIs for everything else.
Automated AI video surveillance shaped for smart cities, police and government, enterprise campuses, retail, healthcare, transport, manufacturing, education and banking.
Built from the ground up for reliability, security and scale. Your video stays in your building, access is controlled to the permission, and Enterprise plans include 24×7 dedicated support.
Encrypted connections end to end, encrypted credential storage, passkey and single sign-on, 46 permissions across 4 roles and a tamper-evident audit trail. Built for mission-critical video.
Visylix never phones home. The whole system, including the AI and the Radha copilot, runs with no internet connection at all, so a fully air-gapped site gets the complete product. Enterprise plans add a 24×7 SLA.
Primary and standby servers with cutover in under 60 seconds, continuous health monitoring and a device-health console. Large camera counts on standard servers, and multi-site federation for the largest estates.
Face recognition, number plate recognition (ANPR), weapons, fire and smoke, PPE, crowds, falls, intrusion, line crossing, tailgating, queues, parking, speed, abandoned objects, camera tampering, anomalies, audio events and more. Every one is built in-house and runs on your own servers. Choose the analytics you need for each camera, billed per camera per month.
Visylix is designed for the operating conditions of manufacturing floors, smart city command centres, hospitals, and transport hubs: continuous streams, on-premise data custody, and analytics that run without an operator watching every feed.
22
In-House AI Analytics
8
Industry Solutions
13
Streaming Protocols
55
Interface Languages
Start your free trial in minutes, no sales call required. Run Visylix on your own servers and grow from one site to a city-wide estate, with unlimited cameras on the VMS licence.
The questions we are asked most. Hundreds more are answered on the full FAQ page.
Visylix is automated AI video surveillance, built in India by Aptibit Technologies. Instead of recording footage for someone to review later, it watches every camera with 22 AI analytics, decides what matters with Sudarshan rules, and acts on its own: alerts, sirens, door-strike relays, talkdown through the camera speaker and escalation when nobody responds. Radha, the built-in AI copilot, runs the whole system in plain language. Visylix is a complete video management system (VMS) with video walls, maps, an alarm console, PTZ autotracking, sealed evidence export and sub-second live view, and it replaces traditional CCTV, DVR and NVR software.
Starter and Pro cover unlimited cameras on the VMS licence, so adding cameras never changes the plan price. Each server carries large camera counts, and multi-site federation joins sites into one estate. We size the hardware with you for your resolution, recording and AI needs, and the installer tunes itself to the server it runs on.
Yes. Visylix runs entirely on your own infrastructure, including the AI analytics and the Radha copilot, and it works fully air-gapped with an offline licence. Your video, face watchlists and Radha conversations never leave your building. You can also run it at the edge, in a hybrid setup or in your own cloud account.
Visylix offers 22 AI analytics, all built in-house. People: face recognition with watchlists and unknown-person alerts, person tracking, unique person counting, crowd detection, fall and running detection, tailgating, queue and wait time, and heat maps. Vehicles: number plate recognition (ANPR) with allow, block and watch lists and exact plate journeys, parking occupancy and speed, wrong-way and U-turn alerts. Safety and security: intrusion, line crossing, weapons, fire and smoke, PPE, abandoned and removed objects, and camera tampering. Scene and sound: object detection across 80 classes, motion, anomaly detection and audio events such as gunshots, glass breaking and screams. Each camera learns its own normal, and marking a false alarm tightens that camera automatically, never below the limit you set. AI analytics are a separate subscription, billed per camera per month on the cameras you choose.
Sudarshan is the part of Visylix that watches and decides. It draws live AI overlays across 12 models on the feed an operator is focused on, keeps the rest of the video wall clean with severity rings and an alert rail, and lets you arm a watch simply by describing it. Its rules engine combines conditions, time sequences and access-control events, previews a rule against real history before you arm it, and scores every alert so the noisy ones can be retuned in one click. Radha is the part you talk to: an on-premise AI copilot with 111 tools that answers in English, Hindi and Hinglish. Ask it to find a red car from last night, trace a plate across cameras, start a PTZ tour, enrol a face, write an automation or brief the next shift. High-impact actions always ask for confirmation, and every action is audit-logged.
Visylix starts with a free 7-day trial on 1 camera. Starter is $49 a month ($539 a year) and Pro is $99 a month ($1,089 a year). In India, Starter is ₹4,999 a month (₹54,989 a year) and Pro is ₹9,999 a month (₹1,09,989 a year), plus GST. Both cover unlimited cameras on the VMS licence. AI analytics are a separate subscription, billed per camera per month on the cameras you choose; talk to us for AI pricing. Enterprise is priced to your estate.