Frequently Asked Questions

Yes. Operators recall presets in one click, run VMS-driven patrol tours, and build or edit tours themselves in the tour editor. Radha can also send a camera to a preset or start a tour by name.

Yes. Active Directory and LDAP sign-in are configured from Settings, alongside SAML and OIDC single sign-on, and every login is audit-logged.

Yes. Tiled inference splits a high-resolution frame into overlapping tiles and runs detection on each, so small, distant objects that a single downscaled pass would miss are still found.

Yes. Floor-plan maps show live camera status and alarm badges, a geographic map view runs on your own supplied tiles, an operator console handles acknowledging, escalating and closing incidents, and an in-product device-health console flags failing cameras and storage runway.

Yes. A dry-run test shows exactly what a rule’s actions would do if it fired right now, with no side effects, including which physical action is held because the rule is not yet armed. Every real fire is also traced, so you can see afterwards precisely which actions ran, were held, or were skipped.

Yes. Ask an open question like "what happened at the loading dock this afternoon?" and Radha runs an agentic multi-step investigation: it autonomously chains person sightings, AI detections, system events, alerts, and video search into findings plus a time-sorted evidence timeline, and can turn the result into a structured, printable incident report.

Yes. Radha performs semantic video search from plain descriptions — "find a man in a red jacket near the entrance" or "was there a white van this morning" — and returns ranked, time-stamped matches you can play instantly. The search runs on an on-premise vision-language index; no footage or query ever leaves your network.

Yes. Radha can look at any live camera or snapshot and describe the scene using an on-premise vision model ("what is happening on the office camera right now?"), and accepts spoken questions through fully local speech-to-text. Both run on your hardware — no cloud vision or speech APIs are ever used.

Yes, and much of it happens automatically. When an operator marks an alert as a false alarm, that camera's threshold for PPE, object, fire and weapon detection tightens automatically, within fixed bounds and never below the operator's own setting, with a single switch to turn it off. Acknowledging or dismissing an alert with a one-tap reason (false alarm, fog or steam, known person, duplicate) also feeds the effectiveness board, which proposes bounded retunes for noisy sources that an operator approves in one click. Separately, motion, audio, crowd, and camera-tampering detection build per-camera scene baselines and anomaly detection trains per camera. Shipped shipped model weights never retrain on site, every adaptation is bounded, reversible, runs on-premise, and survives restarts, nothing leaves your building, and we do not claim a false-positive reduction figure.

Yes. A rule can cross a video condition with an access-control event from your ACS. For example, a person detected on a camera correlated with a denied badge at the same door surfaces a likely forced-entry or tailgating attempt as a single alert on one wall. Sudarshan reuses the door-to-camera mapping already provided by the access-control integration, and the correlation runs on-premise.

Yes, as a v1 cross-camera appearance search. Pick a person from any detection and Visylix finds appearance matches on your other cameras within a time window and assembles a chronological journey with travel times between cameras. It is appearance-based (visual similarity) and returns candidates for operator review, not automatic identity assertions. Operators confirm each candidate in one click. Vehicle plate journeys are exact, since a plate is a precise identifier.

Yes. Rules combine conditions with ANY (OR) or ALL (AND); an ALL rule fires only when every condition co-occurs on the same camera inside a time window. Actions include glowing the tile, notifying a channel, starting a recording, and calling a webhook, SIEM, PSIM or MQTT. Physical actions (siren, access-control request, talk-down) are held disarmed until two operators arm the rule; once armed, a siren or door request raises a critical dispatch to responders and talk-down plays through the camera speaker, so the alert always fires but the physical action requires a deliberate human decision first.

Yes. Radha learns each site's vocabulary: tell her once to "call this camera the gate cam" and every future conversation understands the nickname. She remembers per-user preferences, and thumbs-up/down feedback and corrections are kept on-box as an evaluation record your team controls. All learning is stored on your own hardware and never leaves the deployment.

No. Radha runs entirely on premise: the purpose-built on-premise language model, the vision model, and speech-to-text all execute on your hardware. Radha has no internet access by design and works in fully air-gapped environments, so it suits government facilities, defense installations, and classified networks without modification.

No. All inference, watch evaluation, and overlay assembly run inside your own infrastructure. No frames, detections, or watches are sent to any external service, so Sudarshan works in fully air-gapped, on-premise deployments without modification. Overlay delivery is sub-second, and overlays are fanned out server-side so one camera's data reaches every watching operator at once.

Yes, per camera, on-premise and within fixed bounds. The per-camera anomaly model trains from the visual embeddings of that scene, and camera tamper, motion, crowd, and audio detection each build per-camera baselines. When an operator marks an alert as a false alarm, that camera's threshold for PPE, object, fire and weapon detection tightens automatically, never below the operator's own setting, with a single switch to turn it off. The effectiveness board also proposes retunes an operator approves. Shipped shipped model weights never retrain on site, nothing leaves the premises, and every adaptation is reversible and survives restarts. We do not claim a false-positive reduction figure.

Yes. Line Crossing Detection produces in, out, and net people counts per line with hourly breakdowns on a dashboard, not just individual crossing events, the directional counting product that retail and facilities teams expect.

Yes. Queue and Wait-Time Analytics reports the current line length, the live service rate (how fast the queue is actually clearing), and a predicted wait time for someone joining the queue now, computed from that clearing rate, alongside average and maximum wait. It alerts when a line grows past your threshold so staff can open another counter before customers leave.

No. Visylix AI is fully on-premise. All inference, learning, scene baselines, face templates, and cross-camera search run on your own hardware and work in air-gapped deployments. Your video and model adaptation never leave your building.

Radha enforces role-based permissions, per-camera access restrictions, and licence entitlements on every tool call, and writes every AI action to the audit trail. Radha has no delete tools at all, so nothing can be removed from the system through her. The 13 high-impact actions (arming rules, restarting services, bulk camera operations) require an explicit user confirmation that names the exact target and is enforced server-side — the model cannot approve its own actions. Person searches based on race, ethnicity, or religion are refused, and prompt-injection attempts are blocked and audited.

Every overlay frame is filtered on the server against your licence before it is sent. Unlicensed model data never reaches the browser, so the boundary cannot be bypassed from the client or by inspecting network traffic. Overlays are also scoped per camera and appear only on cameras explicitly configured for AI. Live Intelligence is included with Visylix AI licences.

Sudarshan scores every alert for dispatch priority and measures which alert sources actually earn an operator response, then suggests retunes for the noisy ones. This turns a firehose of alerts into a ranked, actionable queue. The scoring is inspired by ISA-18.2 alarm rationalization and AVS-01 dispatch scoring, and is designed to help operators, not to certify compliance.

The classic failure in rule engines is a rule that looks configured but never fires because the condition it references drifted apart from what the engine actually evaluates. Sudarshan makes that impossible by construction: the builder only offers a condition when the AI model behind it is licensed and enabled on every selected camera, the condition vocabulary is the exact set the engine evaluates and is pinned by contract tests, and a rule that has never fired, has an unarmed physical action, or references a now-unlicensed model is surfaced on a health board. A dead rule is caught, never hidden.

By default Visylix Tailgating Detection counts how many people pass through a secured door during an access event, from video alone, and flags group crossings. Opt in to badge-to-body correlation with a supported access-control system (HID, LenelS2, AMAG, Brivo, Genea) and Visylix compares how many people crossed the door with how many badges were presented, alerting only when more people entered than credentials, which is the true tailgating signal.

Most systems either draw boxes on every camera until the wall is unreadable, or hide AI behind a separate analytics tab. Sudarshan puts live overlays only where the operator is looking and a quiet, glanceable signal everywhere else. It also treats the live feed as sacred: overlays are a thin translucent layer that never covers, freezes, or degrades the video, and stale overlay data fades out rather than misleading the operator.

Radha commands 109 operational tools covering every Visylix feature: live operations (streams, recordings, snapshots, playback, PTZ), AI answers across all 22 models, semantic video search, multi-step investigations, watches and Sudarshan automation rules with dry-run simulation before they go live, camera fault diagnosis and bulk actions across the fleet, server capacity and sizing questions, NVR recorder playback, PTZ presets and patrol tours, the alarm console with sensors, notifications and talk-down audio, daily briefings and incident reports, forensics and evidence locks, and full administration including camera onboarding, user management, face and plate enrollment, zones, sites and floor plans, integrations, licensing, and service restarts. She has no delete tools at all, so nothing can be removed from the system through her.

Sudarshan supports six live watch condition types: specific faces, specific plates, plate patterns, an object class inside a named zone (with minimum count and dwell time), line crossings in a chosen direction, and free-text semantic search over what the camera sees. You describe the condition in plain language and Sudarshan arms a live watch that flags the exact moment it happens on the wall.

Sudarshan Rules is a visual automation builder inside Sudarshan. Where a watch flags a moment for an operator, a rule acts on it automatically, around the clock. You build a rule as IF a condition, WHEN a schedule, THEN an action: for example, IF a person is detected on the loading dock, WHEN it is after 22:00, THEN glow the tile, notify the guard, and start a recording. It is designed for the people who run the video wall, with no scripting or query language.

Sudarshan is Visylix Live Intelligence: a real-time AI layer over your video wall that scales detail to operator attention. The feed an operator is focused on gets rich, live overlays (bounding boxes, tracks, faces, plates, zones, lines), while every other tile stays clean and shows only a signal when something needs attention. It also lets operators watch for anything in plain language and delivers alerts in context. It is named for the Sudarshan Chakra, the ever-vigilant, all-seeing discus.

Cloud VMS vendors run their AI assistants in their cloud; on-premise vendors ship search boxes at most. Radha is a complete copilot (conversational control, semantic search, agentic investigation, vision, voice, reports, and administration) running 100% on premise. It sits inside a 55-language interface, remembers the site vocabulary your team teaches her, and is governed by confirmation gates, query moderation, and a full audit trail.

Sudarshan draws live bounding boxes for twelve models on the focused feed: object detection, person tracking, face recognition, ANPR plate reading, pose estimation, PPE hard-hat compliance, intrusion zones, line crossing, crowd counting, fire and smoke, weapon (gun and knife), and abandoned object. Weapon boxes are shown only after temporal confirmation, so a single-frame false positive never flashes on the feed. Models that have no bounding box, such as acoustic events (gunshot, glass-break, scream), scene-level anomaly detection, camera tampering, and tailgating, still signal on the same wall through the alert lane with a tile ring, a rail entry, and a one-click drill-in.

Yes. Every redaction is an async job queued in the engine. You POST a job, get a job ID, and poll the status endpoint for progress percentage and estimated time remaining. Cancel at any time, partially-rendered output is discarded. For long investigations, you can queue dozens of jobs and the engine chews through them in parallel based on available GPU / CPU resources.

Yes. Every share link can be revoked at any time from the admin console, instantaneously invalidates the token, blocks further access, and writes a revocation event into the audit log. Each share can also be capped at a maximum number of views, and every access is written to the audit log.

Yes. Visylix is fully air-gap compatible and can operate without any internet connection after initial installation. All AI inference, video processing, and management functions run locally with no external dependencies.

Yes, and the deployment model is the reason. Because Visylix runs as a Docker image on infrastructure you control, including fully air-gapped, personal data never has to leave your premises for processing. That removes the cross-border transfer question rather than mitigating it. Your obligations around lawful basis, access control, retention and breach notification remain yours regardless of where the software runs.

Yes. Because Visylix is vendor-neutral and speaks ONVIF plus 13 streaming protocols, you can stage a migration rather than replace an entire estate at once. Compliant cameras can be onboarded alongside existing units, and NVR forensic access across 15+ brands including OEMs lets you keep reaching historical footage on legacy recorders during the transition.

Yes, per user: once a user enrols TOTP or a passkey and enables enforcement, every login requires the second factor, with 5 failed attempts triggering a 15 minute lockout. For an organisation-wide policy, pair Visylix with SSO so your IdP enforces MFA for everyone.

Yes. SSO-provisioned users have no local password, so they sign in only through your IdP (SAML 2.0, OIDC, or Active Directory / LDAP), with group-to-role mapping and just-in-time provisioning. A configurable "break-glass" local admin account is preserved for disaster recovery (lockout protection). Break-glass logins are heavily audit-logged and can trigger alerts to a SOC.

Yes. WebAuthn / FIDO2 authentication is performed entirely between the browser and the Visylix backend, no external identity service required. Passkeys work in fully air-gapped deployments as long as the browser is modern (Chrome 109+, Safari 16.4+, Firefox 119+, Edge 109+). Platform authenticators like iCloud Keychain sync across your own devices; hardware keys like YubiKey are self-contained.

Yes. Every action in Visylix is logged with a timestamp, IP address, user agent, and detailed action information. The comprehensive audit trail covers user logins, configuration changes, evidence access, and all administrative operations.

Yes. Visylix supports optional per-user two-factor authentication (2FA) for an additional layer of security beyond password-based login. 2FA can be enabled or disabled on a per-user basis by administrators.

Yes. Visylix provides four built-in roles: Super Admin, Admin, Operator, and Viewer. The roles map onto 46 granular permissions covering live view, recordings, camera configuration, AI configuration, and system settings.

Not on an on-premise deployment. 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 for the assistant; Visylix does not, which is what makes air-gapped operation possible.

No, and any vendor implying otherwise is overselling. Section 889 is primarily about equipment. A compliant VMS is a necessary component but not sufficient on its own. Your cameras, NVRs, encoders, and network gear all sit in scope. What a VMS can do is avoid adding new exposure, support you in operating compliant hardware, and let you keep using existing non-covered cameras rather than forcing a rip-and-replace.

An operator generates a share link for a locked clip. The link carries a short-lived, cryptographically-signed token. Recipient opens the link, no Visylix account required, and can view (or download, depending on policy) the clip until the token expires. Every access is logged: timestamp, source IP, user agent, referrer. Starter tier supports one recipient per share. Pro supports multiple named recipients with per-recipient expiry and per-recipient access logs.

Face recognition is available but should be treated as a policy decision rather than a default. Several jurisdictions, including Illinois under BIPA and the EU under GDPR Article 9, treat biometric data as a special category with consent and retention requirements. In deployments where biometrics are not required, intrusion, loitering, line crossing and crowd analytics answer most operational questions without processing biometric data at all.

Visylix supports dynamic privacy masks that black out configurable regions in both live view and recordings. Masks are applied server-side before video reaches any viewer, ensuring privacy enforcement cannot be bypassed. Each camera can have its own mask configuration.

You select a clip and pick a redaction preset, faces, license plates, or both. The Visylix engine decodes the clip (H.264 or H.265 input), runs the relevant detector (Face AI for faces, ANPR for plates), blurs each detected face or plate as a rectangle, blurs any custom polygon regions you draw exactly to their shape, then re-encodes the export to H.264 MP4. The whole pipeline runs in a single process, no external transcoder subprocess, no Python video bridge, and each redaction runs as a background job you can poll and cancel.

Five failed MFA attempts within 15 minutes lock the account for 15 minutes. After lockout, the counter resets but the source IP is tracked for anomaly detection. The lockout is keyed to both the user and the source IP, so one attacker cannot keep guessing from the same address. All failures are audit-logged with the attempted factor type, source IP, and user agent.

When a user signs in via SSO, Visylix reads the SAML attribute or OIDC claim you designate (typically a "groups" or "roles" claim). You map those IdP group names to Visylix RBAC roles. A user in IdP group "Security-Operators" automatically receives the Visylix "operator" role on each login. Roles update on every sign-in, so removing a user from an IdP group immediately revokes their VMS permissions on next login.

Not long: SSO is a short form in Settings, and Visylix validates your IdP's discovery document before you switch it on. You paste your IdP metadata URL into the Visylix admin console, map IdP groups to VMS roles (admin, operator, viewer, custom), and enable Just-In-Time (JIT) provisioning. No manual user sync, no SCIM setup required for the basic flow.

Detected faces and plates are blurred as rectangles, following the bounding box the AI model returns for the moments that face or plate is on screen. For whiteboards, screens or documents, you can add custom polygon regions, which are blurred exactly to the shape you draw using ray-casting point-in-polygon tests, so nothing outside the shape is touched. Redaction applies to exports.

Visylix gives you the tools GDPR asks for. It supports on-premise deployment for complete data sovereignty with no cloud dependency required. Configurable retention policies with automatic cleanup ensure data minimization. Dynamic privacy masking protects personal data in video streams.

Visylix is software and is camera-agnostic, so NDAA Section 889 questions attach to the cameras and recorders you deploy, not to the VMS. Visylix does not require or bundle any specific camera hardware, which means it can be run entirely on NDAA-conformant equipment of your choosing.

No. Visylix holds no formal security certification today, and we will not imply otherwise. What it provides are the technical controls those frameworks assess: on-premise or air-gapped deployment, encryption in transit, role-based access control with 46 permissions, configurable retention, evidence lock with SHA-256 hashing, and full audit logging of every operator action. Formal certification is on the roadmap.

The SHA-256 sealed at lock time and the audit trail provide a chain-of-custody record. Visylix exports a self-contained, offline player that plays the clip, recomputes its SHA-256 in the viewer's browser against the sealed hash, and shows a human-readable timeline of custody events. Admissibility is ultimately determined by the court and jurisdiction, but Visylix provides the technical artifacts prosecutors and defense attorneys will ask for.

No. The redaction re-encodes the video with blurred regions baked into the compressed bitstream. There is no reverse transform. This is intentional, redaction is only valid for compliance and evidence sharing if it cannot be undone. The SHA-256 hash on the output file provides tamper-evidence: if a single byte changes after redaction, the hash changes, and downstream systems can detect it.

Three classes. Time-based OTP (TOTP) via Google Authenticator, Authy, 1Password, or any RFC 6238 app, available on Starter and above. WebAuthn / FIDO2 passkeys via hardware keys (YubiKey, SoloKey, Feitian) and platform authenticators (iCloud Keychain, Windows Hello, Android biometric, Google Password Manager, 1Password), available on Pro and above. Backup recovery codes issued when MFA is first enabled, 10 codes, hashed with Argon2id, single-use.

Every view, export and configuration change is written to an audit log against the operator who performed it. Recordings can be hashed with SHA-256 and locked so retention rules cannot delete them while an investigation is open, which supports a chain-of-custody record. Logs can be streamed to a SIEM such as Splunk, Elastic or Datadog via webhook or syslog.

It prevents a clip from being auto-deleted by the retention policy. When you lock a clip, Visylix marks it as "held", the nightly retention job skips it even when it is older than the configured retention window. On Pro and above, locking also computes a SHA-256 hash of the clip bytes and writes it to the clip metadata. Any later modification changes the hash, proving tampering. The hash, plus every subsequent access, forms a chain-of-custody trail visible in the admin audit log.

Visylix enforces TLS 1.2+ for all external connections with TLS 1.0/1.1 disabled. WebRTC media streams use DTLS-SRTP with ECDSA P-256 certificates. All stored credentials are encrypted with Fernet (AES-128-CBC with HMAC-SHA256 authentication). HTTPS is enforced across the dashboard, API, and all management interfaces.

Users receive 10 single-use recovery codes when MFA is first enabled. Each code is hashed with Argon2id and stored one-way, Visylix cannot recover lost codes and neither can we. Admins can also reset MFA for a user, which triggers a full re-enrollment on next login. Recovery-code use and every admin reset are written to the audit log.

Section 889 of the US National Defense Authorization Act restricts federal agencies, and anyone contracting with them, from procuring or using certain Chinese-manufactured telecommunications and video surveillance equipment. Part A bans the federal government from buying it. Part B, effective August 2020, is the far-reaching one: it bars the government from contracting with any entity that uses that equipment anywhere in its operations, even on unrelated systems and even when the government is not paying for it.

Evidence management in Visylix includes timestamped bookmarks with headlines, descriptions, custom colours, and tags. Evidence locks make recordings tamper-evident with SHA-256 integrity verification. Full chain of custody tracks who locked evidence, when, and why.

Starter gets basic Evidence Lock, a clip marked locked cannot be auto-deleted. Pro adds SHA-256 chain-of-custody (each recording sealed with a SHA-256 hash at lock time), time-limited external share links, and the ability to designate multiple recipients with custom per-recipient expiry. Pro also exports a self-contained, offline-verifiable evidence player that recomputes the clip's SHA-256 in the recipient's browser, and includes the SDK and full API for case-management integrations.

Wherever you put it. Visylix does not host your video. It runs on your servers, your virtual machines or your own cloud account, and recordings are written to storage you control, with retention configurable per camera and per site. There is no Visylix-operated cloud holding customer footage.

Any IdP that speaks SAML 2.0 or OpenID Connect (OIDC). We have tested turn-key flows with Okta, Azure AD / Microsoft Entra ID, Google Workspace, Auth0, OneLogin, Keycloak, PingFederate, JumpCloud, and Duo SSO. If you can share a SAML metadata URL or an OIDC discovery document, we can federate with you.

Privacy redaction is included on Enterprise: faces (with Face AI) and plates (with ANPR) are redacted together, plus polygon editing tools for manual redaction of arbitrary regions (useful for exposed whiteboards, screens, or sensitive documents captured in frame).

The statute names Huawei, ZTE, Hytera, Hangzhou Hikvision, and Dahua Technology, along with their subsidiaries and affiliates. The affiliate and OEM dimension is what catches most buyers: a camera sold under a different brand may still be a rebadged unit from a covered manufacturer. Checking the brand on the housing is not sufficient diligence.

Federal agencies, federal contractors and subcontractors at any tier, grant recipients using federal funds, and a growing number of state governments, universities, healthcare systems, and critical-infrastructure operators that have adopted the standard voluntarily or through their own procurement rules. Many enterprises now apply it as a general supply-chain security baseline regardless of whether they hold a federal contract.

Yes. Air-gapped deployment is supported, and it includes the AI. All 22 analytics models and the Radha copilot execute locally, so no video, metadata, or query text is sent anywhere for processing. Many platforms that claim on-premise deployment still call out to a cloud service for inference or for the assistant; Visylix does not.

Yes, through multi-cluster federation. Independent deployments across sites can be searched and viewed from one operations centre while each site keeps its own recordings and retention policy locally. Video stays at the source rather than being centralised into one archive.

Yes. Federated live view uses signed WebRTC URLs. Operator clicks a remote camera, Visylix mints a short-lived (5-minute) HMAC-signed URL at the owning peer, and the browser establishes a direct WebRTC session to that peer's engine. Video never round-trips through the federation coordinator, it flows directly from the source site to the operator's browser over TURN relay.

No. Visylix sells software, not hardware or infrastructure. You run it on your own servers, your own virtual machines, or your own cloud account. This is a deliberate difference from cloud VMS vendors that bundle proprietary cameras or require their hosting.

Yes. Evidence locks, SHA-256 hashes, audit logs and RBAC state replicate synchronously to the standby's PostgreSQL database, so chain of custody survives the switch, and the audit log captures the failover itself as a timestamped event.

Yes. Primary and standby nodes can both live inside your air-gapped network. Replication uses private IP and does not require any internet connectivity. License validation also supports fully offline operation.

No. Pricing is per plan rather than 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.

Each peer has a health check, and a federated search waits at most 30 seconds per peer, returning every site that answered plus a named list of peers that did not, so one slow WAN link never blocks the result.

Visylix HA pairs a primary server with a standby. On primary failure an operator initiates cutover (RTO under 60s) via DNS or L4 load balancer redirection, and leader election hands background work to the surviving node.

When an operator runs a query, "find all red cars between 2 and 4 PM yesterday across every site", Visylix scatters the request to every registered peer in parallel. Each peer executes the search locally against its own cameras and database, returns matching clips and metadata, and Visylix gathers results into a unified feed ranked by timestamp or relevance. 10 sites respond in the time of the slowest one, not the sum of all.

As a Docker image that runs on infrastructure you already control. There are five supported modes: cloud, on-premise, edge, hybrid, and fully air-gapped. Because it is a container rather than an appliance, the same build runs in all five, and moving between them does not mean changing product.

Each federated peer gets a unique HMAC-SHA256 signing secret. Every federation token is HMAC-signed with that peer's own secret, bound to the user and the exact operation, and expires within 60 seconds, so a captured token cannot be replayed against another endpoint.

RTO (recovery time objective) is under 60 seconds. HA is primary/standby with an operator-initiated cutover via DNS or L4 load balancer, and the exact time depends on your DNS TTL or load balancer health-check interval. RPO (recovery point objective) is zero for evidence locks, audit records and RBAC state, because PostgreSQL replicates them synchronously to the standby, and up to about 5 seconds of footage per stream is at risk during a cutover. Configurable health checks monitor CPU, memory, disk, and service liveness.

Running recording and analytics locally at a remote site, such as a depot, branch, or plant, with only alerts and metadata travelling back to a central system. It suits sites with limited or unreliable bandwidth, since full video does not need to cross the link continuously.

Federation lets multiple independent Visylix VMS instances appear as one to the operator. You search across every site from a single box, live-view any camera from any site in one click, and receive events from every peer in one unified feed, but video stays at the source. No central archive, no bandwidth mega-pipe, no single point of failure.

Not in a native app, because there is no mobile app today. The Visylix web interface is responsive, so responders can acknowledge alerts from a phone browser, and critical alerts also reach them by SMS, email or webhook. You can acknowledge alerts one at a time or in bulk with a single tap, so responders in the field can triage and clear alerts without returning to a workstation.

Yes. Configure any number of "pages", each is a named collection of cameras with its own grid layout. The wall cycles through pages on a configurable interval (default 30 seconds, range 5-600), and pins to any page holding an unacknowledged critical alert until it is acknowledged.

Yes. Alerts carry a severity (critical, high, medium, low) derived from AI confidence, zone configuration, and event type. Per-user preferences and quiet hours decide which severities reach each person, and rules can send critical events to SMS, email, webhook or MQTT.

Yes. Outbound PSIM adapters push incidents to Genetec Mission Control, Bosch BIS, and any platform that accepts an HMAC-signed JSON webhook. Each event carries a deep-link back to the Visylix forensic view so operators can click through to the video without leaving the PSIM.

Yes. NVR forensic access covers 15+ brands including OEM rebrands, so historical recordings on existing recorders stay searchable while new recording moves to Visylix. This is what usually makes a migration incremental rather than a rip and replace, since you are changing what records going forward without losing access to what is already stored.

Yes. Retention policies are configurable per camera and per site, so an entrance can be kept longer than a general corridor on the same system. This matters where different areas or regions carry different legal requirements for how long footage may be held.

This is a design question rather than a shipped one, because the apps are not available yet. Aptibit operates no managed cloud service of any kind today. Alerts available now reach you through email, HMAC-signed webhooks (Slack, Teams, PagerDuty), MQTT, SMS, and in-app notifications, none of which depend on a Visylix-operated cloud.

Yes. Pro tier and above include official Python and TypeScript SDKs auto-generated from the OpenAPI spec. Both cover all 450+ REST endpoints: live stream control, recording management, AI detection queries, camera provisioning, RBAC operations, webhook subscriptions. All types, request shapes, and response shapes are strongly typed.

Yes. Visylix ships with a native MQTT broker for IoT sensor fusion (temperature, smoke, motion, ACS, SCADA). Alerts and detections publish to configurable topics, and external sensors can publish events to Visylix for correlated video-and-sensor timelines.

Yes. Enterprise tiers recordings older than a set age to your own S3-compatible bucket: AWS S3, Cloudflare R2, Wasabi, MinIO, Ceph or Cloudian. Archived footage remains instantly searchable via metadata and can be pulled back on-demand.

No. The streaming engine is first-party rather than a wrapper around FFmpeg, SRS, or a similar project. Protocol maturity is stated honestly rather than uniformly: RIST and GB28181 are ingest protocols, GB28181 runs over UDP, and NDI is ingested natively, in process.

No. The engine keeps a GOP cache, so a viewer joining a stream receives the most recent Group of Pictures immediately and starts playing rather than waiting for the camera to emit its next keyframe. On cameras with long keyframe intervals this is the difference between instant playback and several seconds of black.

No. All 22 analytics run against the same decoded frames the engine already produces for live view and recording, on the same node, with no separate analytics server and no per-stream AI fee. Detection zones can be set per camera and per model, so different models can watch different regions of the same feed.

No. Because the stream is stored without re-encoding, recorded footage is identical in quality to what the camera sent. Generational loss only occurs when video is decoded and re-encoded, which Visylix avoids on the recording path. Transcoding is used only when a client requires a different codec or resolution than the camera produces.

Yes. Visylix picks the right stream per tile out of the box. Tiles under 640px (typical 4×4 and denser grids) automatically pull the camera sub-stream, lower bitrate, lower CPU, lower network. The main stream is used for 1×1 views and for the tile under focus mode. This keeps a full 6×6 wall light on bandwidth and CPU.

Opt in to badge-to-body tailgating correlation and Visylix compares how many people crossed a door with how many badges were presented at that door, alerting only when more people entered than credentials, which is the true tailgating (piggyback) signal. It reuses the door-to-camera mapping from the access-control integration and supports HID, LenelS2, AMAG, Brivo and Genea. Without an access-control system, Visylix still performs group-crossing tailgating detection from video alone. The correlation runs on-premise.

Per-user subscription preferences let each operator opt in only to the alert types, cameras, and severities they care about. Quiet-hours scheduling silences non-critical alerts outside working hours. Severity routing delivers critical alerts to oncall rotation while routine events only appear in the dashboard. Acknowledgment tracking prevents duplicate escalation.

Recordings can be hashed with SHA-256 and locked, which prevents retention rules from deleting them while an investigation is open and provides a chain-of-custody record. Every access and export is written to the audit log against the operator who performed it, so it is possible to show who viewed or exported a given clip and when.

Live view uses WebRTC and is sub-second, up to sub-200ms in ideal conditions in internal testing, with a published benchmark planned for Q3 2026. Actual latency depends on the network path between camera, server, and viewer, so it is measured on your own infrastructure during the proof of concept rather than quoted as a fixed number.

Up to 36 cameras per screen in a 6×6 grid. Grid layouts: 1×1, 2×2, 3×3, 4×4, 5×5, 6×6, pick per screen. Each browser window runs its own wall, so a control room can open one per display.

The engine is built for thousands of concurrent streams per node on commodity hardware, but the exact figure depends on resolution, codec, and how many AI models are enabled per camera. There is no published load benchmark yet, so capacity is sized and validated on your own hardware during the proof of concept rather than quoted as a fixed number.

It depends on camera count, resolution, frame rate, codec, and retention period rather than on any fixed per-camera figure. HEVC typically halves the bitrate of H.264 for equivalent quality, so codec choice often has a larger effect on the storage bill than anything else. Sizing is done against your own cameras and retention policy during the proof of concept.

Not today. White-label branding currently covers the web interface (logo, favicon, per-deployment branding, API-managed). A branded mobile app is not offered, since the app itself is not available.

Four channels available today: SMS via Twilio (Enterprise tier), email with inline event thumbnails, HMAC-signed webhooks for custom endpoints, and MQTT for IoT/SCADA integration. Mobile push would arrive with the iOS and Android apps. Each channel is independently subscribable per user.

Bearer token (JWT) for interactive sessions, scoped API keys for automation, and HMAC-signed webhooks for outbound events. All tokens and keys are scopable to RBAC roles and revokable at any time. Webhook receivers can verify the HMAC signature to ensure the event is authentic and un-tampered.

The engine applies adaptive backpressure, dropping non-reference frames first so that keyframes survive. The practical effect is that a struggling viewer sees reduced smoothness rather than a frozen or broken picture, and recording quality is unaffected because recording writes the camera bitstream directly rather than the viewer stream.

Transmuxing repackages an encoded stream from one transport into another without decoding the video pixels. When a client can play the camera's native codec, Visylix transmuxes rather than transcodes, which preserves original image quality and keeps compute cost low. Transcoding runs only when a client genuinely needs a different codec, bitrate, or resolution.

Zero-CPU recording writes the camera bitstream straight to an MP4 file without decoding or re-encoding it. Because no transcoding happens on the recording path, the CPU cost of recording a camera is close to nothing, and the recorded file is a bit-for-bit copy of what the camera produced. Most of a recording server's CPU load in traditional systems comes from transcoding that does not need to happen.

Arrow keys navigate the focus ring between tiles. Enter enters focus mode on the selected tile (main stream, larger, side-by-side with event feed). Escape exits focus mode. Numbers 1-6 switch grid density (1×1 through 6×6). These work without a mouse, designed for SOC operators.

It is a first-party native engine rather than an integration layer over a third-party media server. Each camera is decoded once and the decoded frame is shared by reference across the live-view, recording, and AI paths instead of being copied for each consumer. That single design decision is why the marginal cost of adding one more camera stays low.

When the network is unreliable. SRT and RIST add packet-loss recovery and encryption designed for lossy links such as cellular, satellite, or long-haul internet, where plain RTSP degrades badly. RTSP remains the right choice on a clean local network. Note that RIST support is ingest-only today.

There is no committed date for native apps. Visylix is web-first: live view, alerts, investigation and administration run in any modern browser, desktop or mobile, with nothing to install and no app store in the way.

Visylix pushes security events directly to Splunk HEC, Datadog, and any HTTP collector such as Elastic or QRadar, and emits structured JSON logs any SIEM agent can ship. Every event carries structured fields (camera, zone, AI model, confidence, severity, user, timestamp, source IP) for correlation rules and SOC playbooks.

Native inbound ACS event integrations for HID (Origo/Mercury), LenelS2 OnGuard, AMAG Symmetry, Brivo and Genea/Openpath, plus a generic event format. Door events correlate with camera video and face recognition so badge reads, door holds, and forced-door alarms appear on the same timeline as the corresponding video. Sudarshan rules can require a video condition and a badge event together, and a door request raises a critical dispatch to a responder rather than moving the lock.

H.264 and H.265 (HEVC) for video, and AAC, G.711 and Opus for audio, processed natively by the engine. Streams are passed through without transcoding, so the camera’s own bitstream reaches recording and playback untouched and the CPU stays free for analytics.

No native app is scheduled, so there is no app feature list to publish. Today the responsive web interface gives phone browsers live view over WebRTC, alerts, bookmarks and PTZ.

Thirteen: RTSP, RTMP, HLS and Low-Latency HLS, MPEG-DASH, HTTP-FLV, HTTP-TS, WebRTC using WHEP and WHIP, SRT, RIST (ingest), GB28181, OMT, and native NDI ingest, plus ONVIF for camera discovery and PTZ control. They are implemented in a first-party engine rather than wrapped around an existing open-source media server.

Video Wall is included on the Enterprise plan. It runs in the browser with 1×1 to 6×6 and picture-in-picture layouts, auto-cycle, and critical-alert page pinning. See the pricing page for full tier comparison.

No mainstream browser supports RTSP natively, because browsers standardised on protocols that work over HTTP and survive firewalls and proxies. Any system showing camera video in a browser converts it first. Visylix converts to WebRTC for sub-second live view, or to HLS where broad compatibility matters more than latency.

There is no mobile app today. The responsive web interface works from a phone browser against any Visylix instance you can reach, on-premise or air-gapped. Open your server URL in the browser, sign in with your VMS credentials, and you are ready to go.

Yes, through Sudarshan rules. Conditions can be combined with AND/OR logic and ordered into temporal sequences, so a rule can require a sequence of events before firing rather than a single detection. Rules can be dry-run against past footage first. Actions notify people and systems; they do not directly actuate physical equipment.

Yes. Visylix is a flexible platform that can be customized for any industry beyond the 8 pre-built solutions. Contact our team to discuss your specific requirements and we will build a tailored solution with the right AI models and integrations.

Yes. The unique person counting model holds in-memory appearance embeddings for the length of the counting window and then discards them, which removes double-counting from the footfall figure. That de-duplicated count can be paired with POS transaction data for conversion rate calculation.

It can detect and alert, not physically stop. Detection zones are set per camera and per model, so a press area can require PPE while a walkway does not. Sudarshan rules combine conditions with AND/OR logic and ordered sequences, and actions notify operators and systems. Visylix does not actuate machinery or interlocks.

Yes. Visylix provides classroom and facility occupancy monitoring, parking lot analytics with real-time availability tracking, and heat map analytics for foot traffic patterns. These insights help administrators optimize room scheduling, maintenance, and security patrol routes.

Yes. Visylix provides queue and wait-time analytics with predicted wait, footfall counting, and face watchlists for known offenders or VIPs. These insights help branch managers understand wait times and staffing patterns, using 450+ REST endpoints for integration with banking systems.

Yes. Visylix supports 13 streaming protocols including ONVIF for auto-discovery of existing cameras, RTSP for IP cameras, and GB28181 for China national standard compliance. It provides 450+ REST endpoints for integration with existing city management systems.

Yes. Visylix offers 450+ REST endpoints, IoT MQTT support, and 13 streaming protocols for seamless integration with traffic management systems, toll collection platforms, and fleet tracking software.

Yes. Visylix provides 450+ REST endpoints for integration with hospital information systems, nurse call systems, and visitor management platforms. It supports 13 streaming protocols for connecting existing hospital camera infrastructure.

Yes. Visylix supports multi-site management from a unified dashboard with thousands of streams per node, 450+ REST endpoints for integration, and NVR forensic access across 15+ brands including OEM lines, among them Hikvision, Dahua and Uniview.

Yes. Queue and wait-time analytics and heat-map analytics are two of the 22 in-house AI analytics, and both run on the same streams already being recorded. There is no separate analytics licence and no per-stream AI fee. Detection zones can be set per camera and per model, so a till area and an aisle can be measured differently on the same feed.

Yes. Visylix watches production areas with per-camera anomaly detection, PPE compliance, intrusion zones and line-crossing counts, flags the unusual event nobody wrote a rule for, and triggers real-time alerts via in-app, email, webhook, or WebSocket channels.

Yes. Visylix ships as a Docker image that runs on the bank’s infrastructure, including fully air-gapped environments. All 22 AI analytics and the Radha copilot execute on-premise, so branch footage and any queries about it never leave the bank network. This is the usual deployment mode where data residency rules apply.

Yes. Visylix ships as a Docker image and supports fully air-gapped deployment. All 22 AI analytics and the Radha copilot run on-premise, so no video, metadata, or query text is sent to an external service. This is the deployment mode used where data residency or classified-environment rules apply.

Yes, by licence plate. Plate journeys are exact: the same plate read at different cameras is assembled into one timeline with travel times between reads. Person journeys are also supported but work on appearance rather than identity, so they help follow a route through a terminal without being identity-grade recognition.

Yes. Role-based access control with 46 permissions means a head teacher, a site manager, and a security officer can each see a different subset of cameras and a different set of functions. Every view and export is audit-logged, so it is possible to show who accessed what footage and when.

Yes, through multi-cluster federation. Each store keeps recording locally while a central operations centre monitors all sites, with role-based access control deciding who sees which store. Retention policies are set per site, which matters when different regions have different rules on how long footage may be kept.

Yes. Removed-object detection ships and runs alongside abandoned-object detection, so a tool or part taken from a marked area can raise an alert. Combined with the per-camera anomaly model, which learns each view’s normal pattern and refreshes weekly, this covers both the specific rule and the unusual case nobody wrote a rule for.

Yes. Sudarshan rules combine AND/OR conditions and ordered temporal sequences, so a rule can require several things to happen in a specific order before it fires. Rules can be dry-run against historical footage before going live. Physical actions notify people and systems; they do not actuate hardware such as barriers or signals directly.

Yes. Sudarshan rules can require both a video condition and an access-control condition before firing, so a door opening without a matching badge event can raise an alert while an ordinary authorised entry does not. Conditions can be ordered into sequences, and rules can be dry-run against historical footage before going live.

No. Every paid plan includes unlimited streams, so a chain adding stores or cameras does not pay more in licence fees. Plans start at $49 per month, with Indian pricing from Rs 4,599 per month. Per-camera licensing is the normal model in retail VMS, and it is the reason multi-site rollouts often stall on cost.

Visylix is designed to support SOC 2, ISO 27001, GDPR, HIPAA, and EU AI Act compliance programs. Industry solutions like Healthcare and Banking include additional compliance features such as audit trails, data retention policies, and access control logging.

Yes. Visylix provides SHA-256 evidence lock, retention configurable per camera and site, and a complete audit trail for every video access. It is 100% on-premise with zero cloud dependency and supports air-gapped deployment.

Yes. Visylix is 100% on-premise with zero cloud dependency, supporting air-gapped deployment, and is designed to support GDPR and SOC 2 programs. It includes dynamic privacy masking, RBAC with 46 permissions, configurable retention policies, and comprehensive audit logging. Formal SOC 2 Type II certification is on our roadmap.

Visylix is built for HIPAA-regulated environments: 100% on-premise with encrypted video storage, dynamic privacy masking, RBAC with 46 permissions, configurable retention policies, and comprehensive audit logging. It operates without storing personal biometric data in patient monitoring mode. HIPAA compliance is a shared responsibility that also depends on how your organization deploys and operates the platform.

Yes. Radha produces shift briefings and printable incident reports, and Visylix keeps a comprehensive audit trail of every PPE and zone alert for regulatory audits. All data stays on-premise with air-gapped deployment support for sensitive manufacturing environments.

Yes. Visylix provides ANPR with allow, block and watch lists, exact cross-camera plate journeys, and vehicle counting across ports, warehouses, and distribution centers. NVR forensic access supports 15+ brands including OEMs for reviewing historical footage from existing recorders.

Yes. Object detection, behavior analysis, intrusion detection for restricted areas, and line crossing detection work together for automated loss prevention with real-time alerts via in-app, email, webhook, or WebSocket channels.

It is available in the platform but should be treated as a policy decision, not a default. Many education authorities restrict or prohibit biometric processing of minors, and consent rules differ by jurisdiction. Most school deployments use intrusion, loitering, line crossing, and crowd density instead, which answer the safety questions without processing biometrics.

Yes. It ships as a Docker image and runs with no internet connection at all. All 22 AI analytics and the Radha copilot execute locally, so process footage and any queries about it never leave the plant network. This is a common deployment mode where production data cannot cross the boundary.

No. Visylix uses multi-cluster federation, so several on-premise deployments across districts can be monitored from a central operations centre with role-based access control and per-site retention policies. Each site keeps its own recordings locally, which matters when districts have separate data-custody rules.

Radha runs fully on-premise on the hospital’s own hardware, so nothing it processes is transmitted externally. It has 109 tools available against the camera fleet, the 13 high-impact actions require confirmation, none can delete anything, and every query and action is audit-logged against the operator who made it.

No, when deployed on-premise. Visylix ships as a Docker image that runs on infrastructure the city controls, including fully air-gapped environments. The 22 AI analytics and the Radha copilot all run locally, so no video or metadata is sent to an external service for processing.

PPE detection scores mAP@50 of 0.865 on a frozen 10,315 image evaluation set, the one accuracy figure Visylix publishes. Accuracy on your own site depends on camera angles, lighting, and PPE types, which is why it should be validated during the proof of concept.

Visylix provides AI-powered perimeter intrusion detection with after-hours alerting, floor-plan maps with live camera status, face watchlists for known threats, and secure evidence sharing with law enforcement. It supports thousands of streams per node across multi-building campuses.

Visylix uses pose estimation AI for 24/7 patient fall detection with instant nurse alerts, monitors restricted areas like pharmacies for unauthorized access, and provides emergency department occupancy analytics. All processing is on-premise with zero cloud dependency, which keeps patient video inside your HIPAA boundary.

SRT and RIST are built for exactly this, adding packet-loss recovery and encryption over links that drop. Edge deployment is also supported, so recording and analytics can run locally at a depot or station and only alerts and metadata travel back to the central system.

Visylix builds cross-camera journeys in two forms. Licence plate journeys are exact: the same plate read at multiple cameras is stitched into a single timeline with travel times between points. Person journeys are appearance-based, matching clothing and appearance features across cameras, which is useful for following a route but is not identity-grade recognition and should not be treated as proof of identity.

Visylix provides unique visitor counting with de-duplicated footfall analytics, shopper journey mapping, heat map analytics for merchandise placement optimization, queue management with wait-time estimation, and demographic analytics, all powered by AI models.

Loitering, intrusion, tailgating, abandoned object, and camera tamper detection all run on the same streams already being recorded, with detection zones set per camera and per model. Tamper detection builds a per-camera baseline and adapts to that specific view, which matters for unattended locations where nobody notices a moved or obscured camera.

Visylix detects PPE violations in real time across 8+ gear categories including helmets, safety vests, goggles, and gloves. It enforces PPE per zone around heavy machinery, with detection zones set per camera and per model. Per-camera anomaly and tamper baselines adapt to each view, and when an operator marks a PPE alert as a false alarm, that camera's PPE threshold tightens automatically within fixed bounds, never below the operator's own setting.

Visylix provides ANPR with real-time hot lists and exact cross-camera plate journeys, anomaly, stopped-object and intrusion alerts on highways and in tunnels, passenger flow analytics for terminals and stations, calibrated speed estimation, and vehicle classification and counting, all running on your own infrastructure with no per-camera fees.

Through 450+ REST endpoints, HMAC-signed webhooks, and MQTT for IoT and SCADA. Access control events can be correlated with video, so an alert can require both a badge event and a camera detection before it fires. SSO and MFA are supported for operator authentication.

By plan, not by camera. Every paid plan includes unlimited streams, so adding cameras or sites does not change the licence cost. Plans start at $49 per month, and Enterprise is quoted per deployment. Most enterprise VMS platforms charge per camera or per channel, which is why their cost curve and Visylix’s diverge as an estate grows.

Visylix detects loitering, intrusion, camera tampering and abandoned objects at ATMs in real time. For branches, face recognition identifies known offenders and VIP customers. Tamper detection builds a per-camera baseline, and alerts carry a dispatch score so the effectiveness board can propose retunes for noisy sources.

Visylix combines face recognition correlated with access-control events, perimeter intrusion detection with AI-powered virtual perimeters, after-hours motion detection, PTZ preset recall and patrol tours, and comprehensive audit trails. Model adaptation is per camera and bounded: motion, audio, crowd, and camera-tampering detection build per-camera scene baselines, and anomaly detection trains a per-camera model from the visual embeddings of that scene. When an operator marks an alert as a false alarm, that camera's threshold for PPE, object, fire and weapon detection tightens automatically, within fixed bounds and never below the operator's own setting, with a single switch to turn it off. The effectiveness board also proposes retunes an operator approves. Shipped shipped model weights never retrain on site, and nothing leaves your building. The remaining models are conventional trained models.

Through on-premise processing and privacy redaction. Because Visylix runs as a Docker image on hospital infrastructure, no footage leaves the building for analysis, including AI inference and copilot queries. Redaction can mask faces or regions before footage is exported, and role-based access control with 46 permissions limits who can view which cameras.

Several mechanisms adapt to the store. A per-camera anomaly model learns what normal looks like for that specific view and refreshes weekly, and camera tamper detection builds its own baseline. When an operator marks an alert as a false alarm, that camera's threshold for PPE, object, fire and weapon detection tightens automatically, within fixed bounds and never below the operator's own setting, with a single switch to turn it off. Alerts also carry a dispatch score, and the effectiveness board proposes bounded retunes an operator approves in one click. Visylix does not claim a fixed percentage reduction, because that depends on your scenes.

Visylix provides city-wide camera network integration with centralized management, real-time traffic flow analysis, public safety crowd density monitoring, license plate recognition for parking and toll automation, and 22 AI models, with per-camera scene baselines on motion, audio, crowd, tamper and anomaly (bounded, on-premise, operator-controlled). It is built for thousands of concurrent streams per node with sub-second WebRTC live view, sized on your own hardware during the proof of concept.

Per plan, not per camera. A plant with 600 cameras carries the same licence cost as one with 60, and additional plants can be federated into a central operations view while each keeps recording locally. Plans start at $49 per month, with Enterprise quoted per deployment.

Per plan, not per camera or per branch. Adding cameras or branches does not increase the licence cost, which is the main structural difference from VMS vendors that charge per channel. Plans start at $49 per month, with Enterprise quoted per deployment, and branches can be federated into a central view while each keeps recordings locally.

Retention is configurable per site and per camera, so wards with different regulatory requirements can hold footage for different periods on the same system. Evidence lock prevents specific recordings from being deleted by retention rules while an investigation is open.

Retention is configurable per camera and per site, so entrance cameras can be kept longer than general corridors if policy requires it. Evidence lock holds specific recordings beyond the retention window while an incident is being investigated, so routine deletion does not destroy material that is still needed.

The engine handles thousands of concurrent streams per node in internal testing and scales out by adding nodes, with multi-cluster federation designed toward city-scale estates.

Typically within hours. Visylix installs as a Docker deployment whose auto-config script tunes database pools, workers and buffers to your hardware, and we configure the right analytics and rules for your sector during the proof of concept.

No. Visylix holds no formal security certification today, and we will not claim otherwise. What it provides are the technical controls those frameworks ask for: on-premise or air-gapped deployment, encryption in transit, role-based access control, configurable retention, and full audit logging of every operator action. Certification is on the roadmap.

No. Visylix holds no formal security certification today. It provides the technical controls those frameworks look for, including on-premise deployment, encryption in transit, role-based access control with 46 permissions, configurable retention, evidence lock, and full audit logging. Certification is on the roadmap and we will not claim it before it exists.

Yes. Visylix scales from small K-12 campuses to large university networks. It is 100% on-premise with zero cloud dependency, supports RBAC with 46 permissions for multi-level access control, and provides 450+ REST endpoints for integration with existing school management systems.

Yes. Visylix is built by Aptibit Technologies in Kolkata, India with 55+ languages including 13 Indian languages (Hindi, Bengali, Tamil, Telugu, Kannada, Malayalam, Assamese, Marathi, Gujarati, Punjabi, Rajasthani, Haryanvi, Odia). It supports air-gapped deployment for government facilities and is designed for Smart City Mission India requirements.

Radha runs entirely on-premise and has 109 tools it can execute against the camera fleet. Its 13 high-impact actions require confirmation before running, none of them can delete anything, and every action is written to the audit log with the operator who triggered it. There is no cloud dependency, so copilot queries never leave the building.

Licensing is per plan, not per camera, so a terminal with 400 cameras costs the same in licence terms as one with 40. Plans start at $49 per month and Enterprise is quoted per deployment. Visylix is delivered as a Docker image on your own infrastructure, so there is no hardware purchase from us.

There is no per-camera or per-channel fee. Every paid plan carries unlimited streams, so the cost does not scale with the number of cameras. Plans start at $49 per month, with Indian pricing from Rs 4,599 per month, and Enterprise pricing is quoted per deployment. This is the main structural difference from VMS vendors that licence by camera count.

There is no per-camera fee, so cost does not scale with camera count. Plans start at $49 per month, with Indian pricing from Rs 4,599 per month. A district can federate multiple schools into one central view while each site keeps its own recordings and retention policy locally.

They usually stay. Visylix discovers ONVIF cameras automatically and speaks RTSP directly, so most IP cameras from the last decade connect without replacement. NVR forensic access covers 15+ brands including OEM rebrands, so historical footage on existing recorders stays reachable while new recording moves to Visylix.

Visylix provides AI-powered video analytics solutions for 8 key industries: Smart Cities, Enterprise Security, Retail Analytics, Transportation, Healthcare, Manufacturing, Education, and Banking & Finance. Each solution is purpose-built with industry-specific AI models and workflows.

Rules vary by country and state, but the consistent principle is that cameras are acceptable in common areas such as corridors, entrances, grounds, and car parks, and not in bathrooms, changing rooms, or other private spaces. Many jurisdictions also require signage and a documented retention period. Check your local education authority rules before installing, and treat this as general guidance rather than legal advice.

Visylix supports 13 natively. RTSP and ONVIF cover fixed IP cameras, SRT and RIST carry video reliably over unreliable links such as cellular or long-haul internet, GB28181 covers the Chinese national standard used in some transport equipment, and WebRTC delivers sub-second live view in a browser for control-room operators.

In most cases yes. Visylix auto-discovers ONVIF cameras and connects to RTSP streams directly, so it is camera-agnostic rather than tied to specific hardware. Existing NVRs from 15+ brands remain reachable for historical footage, so a rollout does not require replacing recorders on day one.

Yes. Visylix implements client-side ONVIF Profile S, T and G operations and accepts any RTSP source, so your existing cameras work without changes. Our team assists with migrating recording configurations, user permissions, and layout setups. Since both platforms support ONVIF, the transition is straightforward.

Yes. Visylix implements client-side ONVIF Profile S, T and G operations and accepts any RTSP source, so your existing cameras work without changes. Our team helps migrate recording configurations and user permissions. Visylix ingests events natively from HID, LenelS2, AMAG, Brivo and Genea and correlates badge events with video in Sudarshan rules.

If your cameras support ONVIF, yes. Visylix connects to any ONVIF compatible IP camera from any manufacturer. Rhombus cameras are designed to work with Rhombus's own cloud, so those specific units would need to be replaced with ONVIF compatible alternatives over time.

Yes. Visylix works with any ONVIF compatible IP camera from any manufacturer. You can keep your existing Axis, Hikvision, Dahua, Hanwha, Bosch, or any other compatible cameras and add Visylix as your video management and AI analytics platform.

Yes. Visylix fully supports air gapped deployments with zero internet dependency. All video management, 22 AI models, and the Radha AI Copilot with 109 tools, purpose-built on an on-premise language model function completely offline. Visylix also provides NVR forensic access across 6 native NVR protocols covering 15+ brands including OEM lines and 450+ REST endpoints. Visylix's AI and copilot are first-party, so there is no third-party plugin whose licensing or connectivity you must check.

Yes. Visylix supports on premise, cloud, and hybrid deployment options with NVR forensic access across 6 native NVR protocols covering 15+ brands including OEM lines and 450+ REST endpoints. For organizations that need complete data sovereignty, all video data, AI processing, and the Radha AI Copilot with 109 tools, purpose-built on an on-premise language model run entirely within your infrastructure. Spot AI is primarily cloud based and requires internet connectivity.

Yes. Visylix supports fully on premise and air gapped deployments. All video management, recording, streaming, AI processing, and the Radha AI Copilot function without any internet connection. Coram AI requires internet connectivity to access its cloud based analytics and search.

Yes. Visylix supports fully air gapped deployments with zero cloud dependencies. All video management, AI processing, recording, and the Radha AI Copilot function without any internet connection. Verkada requires cloud connectivity to operate.

No. Visylix includes a native recording engine that handles continuous and event based recording, storage management, and video playback. You do not need a separate NVR or recording device. Coram AI typically works alongside your existing NVR infrastructure.

Visylix offers 22 AI models with GPU-accelerated inference, scoped to your plan and all developed in-house with no per-stream AI fee. Model adaptation is per camera and bounded: motion, audio, crowd, and camera-tampering detection build per-camera scene baselines, and anomaly detection trains a per-camera model from the visual embeddings of that scene. When an operator marks an alert as a false alarm, that camera's threshold for PPE, object, fire and weapon detection tightens automatically, within fixed bounds and never below the operator's own setting, with a single switch to turn it off. The effectiveness board also proposes retunes an operator approves. Shipped shipped model weights never retrain on site, and nothing leaves your building. The remaining models are conventional trained models. Visylix also includes the Radha AI Copilot with 109 tools, purpose-built on an on-premise language model for natural language system control, which Verkada does not offer.

Visylix is focused on video management and AI analytics. For access control, Visylix ingests events natively from HID, LenelS2, AMAG, Brivo and Genea and correlates them with video in Sudarshan rules. Visylix also provides NVR forensic access across 6 native NVR protocols covering 15+ brands including OEM lines. This allows you to choose best in class solutions for each function rather than being locked into a single vendor.

No. Visylix includes its own native recording engine, so you do not need a separate NVR. The platform handles recording, storage management, and playback natively. Spot AI typically requires an existing NVR or recording infrastructure to function.

Yes. Visylix supports on premise, cloud, and hybrid deployment options. For organizations that require complete data sovereignty, Visylix runs entirely within your infrastructure with zero cloud dependencies, including the Radha AI Copilot. Visylix also provides NVR forensic access across 6 native NVR protocols covering 15+ brands including OEM lines and 450+ REST endpoints for deep integration.

Yes. Visylix connects to any ONVIF compatible IP camera, covering the same broad range of manufacturers that Nx Witness supports, plus ONVIF vendor device packs for the top five brands, and any camera that serves RTSP.

Yes. Visylix supports any ONVIF compatible IP camera, including models from Axis, Hikvision, Dahua, Hanwha, Bosch, and hundreds of other manufacturers. You are never locked into purchasing cameras from a single vendor.

Visylix offers 22 in-house AI models with GPU-accelerated inference, scoped to your plan: face recognition, ANPR, object detection, pose estimation, crowd detection, PPE detection, heat maps, motion detection, person tracking, unique person counting, intrusion detection, line crossing, camera tampering detection, abandoned object detection, tailgating detection, queue and wait-time analytics, parking occupancy, speed estimation, anomaly detection, fire and smoke detection, weapon detection, and audio event detection. Model adaptation is per camera and bounded: motion, audio, crowd, and camera-tampering detection build per-camera scene baselines, and anomaly detection trains a per-camera model from the visual embeddings of that scene. When an operator marks an alert as a false alarm, that camera's threshold for PPE, object, fire and weapon detection tightens automatically, within fixed bounds and never below the operator's own setting, with a single switch to turn it off. The effectiveness board also proposes retunes an operator approves. Shipped shipped model weights never retrain on site, and nothing leaves your building. The remaining models are conventional trained models. Coram AI offers select AI features focused primarily on search and alerting.

Visylix offers 22 in-house AI models with GPU-accelerated inference, scoped to your plan, covering detection, recognition, and behavioural analytics. Model adaptation is per camera and bounded: motion, audio, crowd, and camera-tampering detection build per-camera scene baselines, and anomaly detection trains a per-camera model from the visual embeddings of that scene. When an operator marks an alert as a false alarm, that camera's threshold for PPE, object, fire and weapon detection tightens automatically, within fixed bounds and never below the operator's own setting, with a single switch to turn it off. The effectiveness board also proposes retunes an operator approves. Shipped shipped model weights never retrain on site, and nothing leaves your building. The remaining models are conventional trained models. Visylix also includes the Radha AI Copilot with 109 tools, purpose-built on an on-premise language model for natural language system control. Visylix runs its analytics on the cameras you already own.

Visylix offers complete on premise deployment where all video data, AI processing, and even the Radha AI Copilot with 109 tools, purpose-built on an on-premise language model run within your infrastructure. No data leaves your network. Visylix also provides NVR forensic access across 6 native NVR protocols covering 15+ brands including OEM lines and 450+ REST endpoints. Verkada requires cloud connectivity and stores video data on their servers, which may not meet strict data sovereignty requirements.

Visylix supports large-scale concurrent streams and thousands of streams per node with 13 native streaming protocols including WebRTC, RTSP, NDI, and RIST. The platform is designed for multi site enterprise deployments with centralized management and site level controls, allowing you to manage thousands of cameras across multiple locations from a single interface.

Visylix offers 22 in-house AI models with GPU-accelerated inference, scoped to your plan, covering detection, recognition, and behavioural analytics. Model adaptation is per camera and bounded: motion, audio, crowd, and camera-tampering detection build per-camera scene baselines, and anomaly detection trains a per-camera model from the visual embeddings of that scene. When an operator marks an alert as a false alarm, that camera's threshold for PPE, object, fire and weapon detection tightens automatically, within fixed bounds and never below the operator's own setting, with a single switch to turn it off. The effectiveness board also proposes retunes an operator approves. Shipped shipped model weights never retrain on site, and nothing leaves your building. The remaining models are conventional trained models. There are no plugins to install, configure, or maintain. Nx Witness relies on third party AI plugins that require separate licensing, configuration, and ongoing maintenance, adding complexity and cost.

Visylix deploys as a Docker image on your own infrastructure with full on premise, cloud, hybrid, and air gap options. All video data, AI processing, and the Radha AI Copilot with 109 tools, purpose-built on an on-premise language model run within your network. Visylix also provides NVR forensic access across 6 native NVR protocols covering 15+ brands including OEM lines and 450+ REST endpoints. Rhombus requires cloud connectivity and stores video data on their servers, which may not meet data sovereignty requirements.

Visylix uses flat rate monthly plans starting at $49 per month with no per camera fees. Milestone charges per camera license, which means costs scale linearly as you add cameras. For deployments with hundreds or thousands of cameras, Visylix can be significantly more cost effective.

Both platforms offer AI powered video search. Visylix goes further with the Radha AI Copilot, which can not only search recordings but also control cameras, manage alerts, configure AI models, and run 109 operational tools through natural language. Spot AI focuses primarily on search and retrieval.

Visylix is a complete video management system with native recording, streaming, camera management, and 22 in-house AI analytics. Spot AI is primarily a cloud analytics layer that connects to your existing cameras and NVR. Visylix replaces your entire VMS stack, while Spot AI augments what you already have.

Yes. Visylix uses flat rate monthly plans starting at $49 per month with no per camera fees. You can scale to thousands of cameras without increasing your per camera cost. Nx Witness uses traditional per camera licenses, meaning costs grow linearly with each camera you add to your system.

Yes. Visylix uses flat rate monthly plans starting at $49 per month with no per camera licensing fees. You can scale from a small deployment to thousands of cameras without your per camera costs increasing. Genetec requires per camera connection licenses, and pricing typically requires a custom quote from an authorized integrator.

Yes. Visylix uses flat rate monthly plans starting at $49 per month with no per camera fees. You can connect unlimited cameras on every paid plan. Rhombus uses a Camera as a Service model that charges per camera, so costs scale linearly as you add cameras to your deployment.

Visylix is designed for enterprise scale from the ground up, supporting large-scale concurrent streams and thousands of streams per node with 13 native streaming protocols including WebRTC, RTSP, NDI, and RIST. The proprietary streaming engine is purpose-built for maximum throughput with minimal resource consumption.

Since Verkada uses proprietary cameras, transitioning requires replacing camera hardware over time. However, you can start by adding ONVIF compatible cameras alongside existing Verkada units and running Visylix in parallel. Our team helps plan a phased migration strategy.

Visylix is a complete video management system with native recording, live streaming, camera management, and 22 in-house AI analytics. Coram AI is primarily a cloud analytics platform that adds AI capabilities to your existing camera system. Visylix replaces your entire video management stack, while Coram AI layers analytics on top of what you have.

Visylix offers 22 in-house AI models with GPU-accelerated inference, scoped to your plan, covering detection, recognition, and behavioural analytics. Model adaptation is per camera and bounded: motion, audio, crowd, and camera-tampering detection build per-camera scene baselines, and anomaly detection trains a per-camera model from the visual embeddings of that scene. When an operator marks an alert as a false alarm, that camera's threshold for PPE, object, fire and weapon detection tightens automatically, within fixed bounds and never below the operator's own setting, with a single switch to turn it off. The effectiveness board also proposes retunes an operator approves. Shipped shipped model weights never retrain on site, and nothing leaves your building. The remaining models are conventional trained models. Genetec relies largely on partner integrations for advanced analytics.

Visylix supports large-scale concurrent streams and thousands of streams per node with 13 native streaming protocols including WebRTC, RTSP, NDI, and RIST. The proprietary streaming engine and on premise deployment option with NVR forensic access across 6 native NVR protocols covering 15+ brands including OEM lines and 450+ REST endpoints make it suitable for large, distributed deployments. Coram AI is well suited for small to medium businesses but does not publish enterprise scale benchmarks.

Visylix is designed for enterprise scale, supporting large-scale concurrent streams and thousands of streams per node with 13 native streaming protocols including WebRTC, RTSP, NDI, and RIST. The architecture is optimized for large deployments with efficient resource utilization. Spot AI is well suited for small to mid size deployments but does not publish comparable scalability figures.

Yes. Visylix is 100% on-premise with zero cloud dependency. It is fully air-gap compatible, the entire system operates without an internet connection after initial installation. All video, AI data, face databases, and Radha AI conversations stay within your network. It supports on-premise, cloud, hybrid, and edge deployments via Docker and Kubernetes.

Yes. Visylix is designed to operate in completely disconnected environments with no internet access. All AI models, the streaming engine, recording system, and Radha AI Copilot run locally. This makes Visylix suitable for government, defense, and classified network deployments.

Yes. Visylix automatically transcodes H.265 (HEVC) to H.264 in real time for browsers and clients that lack HEVC hardware support. Audio transcoding (G.711, AAC → Opus) happens server-side for WebRTC delivery with proper sample-rate conversion and channel remapping, so operators never have to think about codec compatibility.

Yes. Direct SIEM push to Splunk HEC, Datadog, and any HTTP collector such as Elastic or QRadar. PSIM adapters push incidents to Genetec Mission Control, Bosch BIS, and any platform accepting an HMAC-signed JSON webhook, with deep-link-back to Visylix for forensic review.

Yes. Beyond ANPR/ALPR license plate recognition, Visylix classifies vehicles as car, truck, bus or motorcycle and tracks direction of travel, alongside ANPR plate reads with multi-frame confidence voting, so every vehicle detection carries both a plate string and a rich appearance profile for downstream search and filtering.

No. Visylix can be deployed entirely on premise with zero cloud dependencies. All video processing, AI inference, recording, and storage happen on your own infrastructure. No video frames, thumbnails, metadata, or AI results leave your network unless you explicitly configure cloud or hybrid deployment.

Yes. Enterprise deployments include HA with leader election and operator-initiated standby cutover (RTO under 60s via DNS or L4). Multi-site federation lets one operator search, receive events from, and live-view cameras across sites while video stays at the source.

Yes. Visylix can access recordings stored on external NVR devices from 15+ brands including OEM lines, among them Hikvision, Dahua, Uniview, I8H, Bosch and Panasonic natively, plus ONVIF Profile G recorders. Features include time-range forensic search, multi-channel forensic grid, variable-speed playback (0.5x to 8x), recording download with SHA-256 integrity verification, and automatic cross-brand capability detection.

Yes. Fisheye cameras get browser-side virtual PTZ in the focused live view, so an operator can pan and zoom inside a 360° image (Enterprise). Enterprise includes a browser video wall with up to 6×6 grids, auto-cycle, and pinning to critical alerts, and a control room can open one wall per display.

Contact our team through this form or email us at info@aptibit.com. We will schedule a personalized demo, understand your requirements, and provide a deployment plan. A standard install is a single Docker deployment on your own server, validated on your hardware during the proof of concept.

Evidence lock seals each recording with a SHA-256 hash at lock time, and exports a self-contained player that verifies the clip offline. The audit trail logs every view, export and share with timestamp, user and IP. Case sharing (Pro+) supports expiring signed links and view caps, and GDPR right-to-delete workflows that preserve hashes for legal compliance.

The engine includes adaptive backpressure (drops non-reference frames while preserving keyframes for slow subscribers) and a circuit breaker monitoring CPU load in real time, gracefully degrading service at 70%, 85%, and 95% CPU thresholds rather than crashing. New viewers receive the cached Group of Pictures instantly, eliminating the "waiting for keyframe" delay typical of other streaming systems.

Visylix supports GDPR compliance through data minimization, configurable retention policies, right to erasure capabilities, and comprehensive audit logging. On premise deployment ensures data sovereignty by keeping all personal data within your chosen jurisdiction.

Alerts are delivered across four channels simultaneously: SMS (Twilio on Enterprise), email with inline thumbnails, HMAC-signed webhooks for custom endpoints, and MQTT for IoT/SCADA systems, alongside in-app notifications. Per-user subscription preferences, quiet-hours scheduling, severity routing, and acknowledgment tracking ensure alerts reach the right people without noise.

Visylix offers four recording modes: continuous (24/7), scheduled, event-triggered (with pre/post-event buffer), and manual. The engine uses zero-CPU recording, video is written directly from the camera bitstream to faststart MP4 containers without any transcoding. H.265 streams are recorded natively, so HEVC's bitrate advantage carries straight into storage. Recordings include automatic thumbnail generation, configurable per-camera retention, and timeline-based playback up to 8× speed.

Visylix offers 22 production-ready AI models, developed in-house and scoped to your plan: face recognition (large-scale identity matching), license plate recognition (ANPR/ALPR with a global plate OCR model, hot lists and exact cross-camera plate journeys), object detection (80+ classes), person tracking (per-camera tracking plus cross-camera appearance journeys, Re-ID v1), crowd detection, PPE/safety gear detection, heat map analytics, motion detection, pose estimation (17 keypoints, fall detection), unique person counting, intrusion detection, line crossing detection (directional in/out/net counting), camera tampering detection, abandoned object detection, tailgating detection (with opt-in badge-to-body correlation), queue and wait-time analytics (with predicted wait), parking occupancy (with auto slot discovery), speed estimation, anomaly detection (semantic per-camera memory), fire and smoke detection, weapon detection, and audio event detection. Model adaptation is per camera and bounded: motion, audio, crowd, and camera-tampering detection build per-camera scene baselines, and anomaly detection trains a per-camera model from the visual embeddings of that scene. When an operator marks an alert as a false alarm, that camera's threshold for PPE, object, fire and weapon detection tightens automatically, within fixed bounds and never below the operator's own setting, with a single switch to turn it off. The effectiveness board also proposes retunes an operator approves. Shipped shipped model weights never retrain on site, and nothing leaves your building. Detection accuracy depends on camera angle, lighting, and scene, so we validate every model on your own footage during the proof of concept.

There is no per-camera limit and no per-camera licence. The engine is built for thousands of concurrent streams per node on commodity hardware, and the exact per-node capacity depends on resolution, codec, and how many AI models you enable, so we size and validate it on your own hardware during the proof of concept. The auto-tuning installer profiles your hardware and automatically configures optimal database pools, worker counts, buffer sizes, and connection limits.

Visylix is designed for high stream density on commodity hardware, thousands of concurrent streams per node, and it scales out by adding nodes. Multi-cluster federation is designed toward very large, city-scale estates. Exact per-node capacity is validated on your own hardware during the proof of concept.

Yes. Visylix is built with privacy-first principles and supports 55+ languages including RTL scripts (Arabic, Hebrew, Persian, Urdu). It supports data anonymization, configurable retention policies, consent management, audit trails, and encrypted storage. Our platform is built to support your GDPR, SOC 2, ISO 27001, and HIPAA compliance programs.

Visylix offers a free 7-day trial with 1 stream, 1 user, and 1 viewer. Paid plans are $49/month (Starter) and $99/month (Pro); in India ₹4,599 and ₹9,299. Enterprise, at custom pricing, adds AI analytics (any subset up to all 22, priced by the analytics you enable). There is no per-camera licensing, flat-rate pricing regardless of camera count.

Visylix offers 22 in-house AI models with GPU-accelerated inference, scoped to your plan, including face recognition, ANPR, PPE, intrusion, tailgating, queue analytics, parking, fire and smoke, weapon, audio and anomaly detection. Per-camera anomaly and tamper baselines adapt to each view. Custom models can also be trained.

The Enterprise tier includes SSO via SAML 2.0 and OIDC (Okta, Azure AD / Entra ID, Google Workspace, Auth0, OneLogin, PingIdentity, any standards-compliant IdP). MFA is available from Starter tier via TOTP with recovery codes. Pro tier and above support WebAuthn / Passkeys (FIDO2) for passwordless login compatible with Apple Passkeys, Google Password Manager, 1Password, Bitwarden, and hardware keys.

Native inbound ACS integrations for HID, LenelS2 OnGuard, AMAG Symmetry, Brivo and Genea/Openpath. Badge reads, door holds and forced-door alarms appear on the same timeline as the video, and Sudarshan rules can require a badge event and a camera detection together, with unified incident timelines across video and access events.

Enterprise tiers recordings older than a set age to your own S3-compatible storage: AWS S3, Cloudflare R2, Wasabi, MinIO, Ceph or Cloudian. Archived footage remains instantly searchable via metadata and can be pulled back on-demand for investigation.

Unlimited polygonal detection zones and directional virtual tripwires per camera. Zones can activate per-AI-model rules (which models run, which thresholds apply) and per-schedule (active hours). Tripwires count entries and exits separately with direction awareness, driving line-crossing analytics and occupancy tracking.

Visylix enforces TLS 1.2 or 1.3 in transit, DTLS-SRTP for WebRTC, and Fernet authenticated encryption for stored credentials; recordings sit on storage you can encrypt with LUKS or BitLocker. Encryption keys are managed locally on your infrastructure.

Radha is the only VMS copilot we know of that is conversational, action-taking and fully on-premise at once. She is purpose-built on an on-premise language model that runs on your own hardware, and she carries 109 tools: natural-language control of cameras and recordings, answers from every AI analytic you have enabled, semantic video search ("find a man in a red jacket"), agentic multi-step investigations with evidence timelines, live scene description via an on-premise vision model, local voice input, daily briefings, printable incident reports, and full administration. Radha keeps the site-specific vocabulary and preferences your team configures, lives inside a 55-language interface, and is governed by confirmation gates, query moderation, and a full audit trail, all working air-gapped.

Visylix is automated AI based video surveillance, built by Aptibit Technologies. It combines 13 streaming protocols, 450+ REST endpoints, NVR forensic access across 6 native NVR protocols covering 15+ brands including OEM lines, 22 AI models, zero-CPU recording, and Radha, the only VMS copilot we know of that is conversational, action-taking and fully on-premise at once. It connects to your existing IP cameras via RTSP/ONVIF/GB28181/NDI/RIST and more, processes video with AI in real time, and delivers actionable intelligence. 100% on-premise, air-gap compatible.

Visylix is automated AI based video surveillance: the system watches, decides, and acts, rather than just recording for someone to review later. It is an enterprise-grade AI video management system (VMS) built by Aptibit Technologies, combining a proprietary native streaming engine, 22 AI analytics, and Radha, the only VMS copilot we know of that is conversational, action-taking and fully on-premise at once. It supports large-scale concurrent connections with sub-second WebRTC live view, up to sub-200ms in ideal conditions, 13 streaming protocols, 450+ REST endpoints, and serves as a modern alternative to traditional CCTV, DVR, and NVR systems.

Visylix supports 55+ languages. This covers 13 Indian languages (Hindi, Bengali, Tamil, Telugu, Kannada, Malayalam, Assamese, Marathi, Gujarati, Punjabi, Rajasthani, Haryanvi, Odia), RTL scripts (Arabic, Hebrew, Persian/Farsi, Urdu), European locales (Spanish, French, German, Portuguese, Italian, Russian, Polish, Dutch, Ukrainian, Romanian, Czech, Hungarian, Greek, Swedish, Danish, Norwegian, Finnish, Bulgarian, Serbian, Croatian, Slovak), East & Southeast Asian languages (Mandarin Simplified & Traditional, Japanese, Korean, Thai, Vietnamese, Indonesian, Malay, Burmese, Khmer, Lao, Nepali, Sinhala), and Turkish, Swahili, and Amharic. The system detects browser language automatically and each user can set their preferred language independently.

Server-side privacy redaction is applied before footage is released to any recipient. Enterprise includes full redaction (faces, plates and custom regions). Redaction runs on exports and shared clips, with RBAC-based exemption rules so authorized roles (e.g., law enforcement) can view un-redacted footage with full audit logging.

Radha AI Copilot includes multiple prompt injection detection patterns, role based access control for all 109 operational tools, per user rate limiting (10 messages a minute), and comprehensive audit logging. Every user input is validated before tool execution, and operators can only use tools that their assigned role permits.

Visylix supports 13 streaming protocols in a first-party engine, including WebRTC (WHEP/WHIP) with sub-second live view, up to sub-200ms in ideal conditions, RTSP with Digest Auth, RTMP, SRT with AES-256 encryption, HLS with Low-Latency HLS (LL-HLS), HTTP-FLV, ONVIF (Profile S/T/G), GB/T 28181, native NDI ingest for broadcast workflows, RIST ingest for broadcast-grade internet links, and native OMT.

The Visylix blog covers AI video analytics, streaming protocols (WebRTC, RTSP, HLS, SRT), edge computing for surveillance, VMS platform architecture, industry solutions (smart cities, retail, manufacturing), competitor comparisons, migration guides, and cybersecurity best practices for video surveillance systems.

Articles are written by Mr. Falguni Chatterjee, Founder & CEO of Aptibit Technologies, and the Visylix engineering team. Content is based on the engineering experience of building an AI video management platform from the engine up.