Official Platform Documentation

FireVision Documentation

Your comprehensive guide to connecting cameras, configuring AI models, managing emergency workflows, and deploying edge gateways.

1. Getting Started & Instant Provisioning

FireVision transforms your existing CCTV cameras into an autonomous 24/7 AI detection and life-safety system. Getting started requires only three simple steps:

Step 1

Account Creation

Sign up securely via Clerk authentication, verify your email, and select your camera monitoring plan.

Step 2

Connect Cameras

Enter your camera RTSP stream or connect a plug-and-play Secure WireGuard Raspberry Pi Gateway.

Step 3

Automated Activation

Streams are verified and provisioned automatically in real-time with zero manual approval queues.

2. Camera & Gateway Connection Methods

FireVision supports three flexible connectivity architectures to suit any property or industrial site:

Method A: Secure WireGuard Pi Gateway (Recommended)

For maximum security without opening inbound router ports, deploy a pre-configured Raspberry Pi appliance. The Pi establishes an encrypted WireGuard VPN tunnel back to our Manchester AI processing cluster.

Zero Port ForwardingHardware Device ID TrackingAutomated MediaMTX Sync

Method B: Off-Grid Solar & 4G Remote Hubs

Ideal for agricultural land, timber yards, and remote estates without mains electricity or broadband. Packages include solar panels, high-capacity batteries, and 4G SIM routers for self-sustaining perimeter surveillance.

Method C: Direct RTSP Stream with DDNS / Port Forwarding

If using your own router, map external port 2554 (or custom) to your camera's internal RTSP port (554) and supply your static IP or DDNS hostname in the Profile page.

Supported Cameras: Works with all standard ONVIF and RTSP (H.264/H.265) IP cameras including Wansview, Dahua, Hikvision, Reolink, and TP-Link Tapo. Proprietary cloud-locked cameras (Ring, Blink) that lack local RTSP access are not supported.

3. AI Detection Models & Precision Scheduling

FireVision allows per-camera configuration of multiple computer vision models and operational schedules:

🔥 Visual Fire & Smoke Detection

Identifies optical flame and smoke growth signatures within seconds, capturing incidents long before smoke reaches ceiling detectors.

👤 MultiVision: Human Intrusion

Guards restricted yards, warehouse docks, and after-hours perimeters with instant notification on unauthorized entry.

🚗 MultiVision: Vehicle & Fleet

Monitors gate access points, unauthorized parking zones, and delivery bays with vehicle presence classification.

🐾 MultiVision: Wildlife & Livestock

Protects pastures and livestock from predators while distinguishing wildlife movement to eliminate false intrusion alarms.

⏱️ 2-Second Pre-Alarm Rolling Buffer & Cooldown

Our stream engine maintains a continuous 2-second in-memory buffer. When a hazard is detected, the generated video clip includes the 2 seconds *prior* to detection, giving first responders immediate context on how the event started. Intelligent cooldown periods prevent notification spam during ongoing incidents.

4. Multi-Channel Alerts & Emergency Evacuation App

When a verified detection occurs, FireVision executes an automated multi-channel escalation sequence:

  • Automated Voice Phone Calls (Twilio): Immediate audible telephone call dispatched to the account owner and registered delegates.
  • Instant SMS Notifications: Text alerts with direct links to review captured event footage in under 10 seconds.
  • Rich Email Alerts: Comprehensive emails containing camera location metadata, timestamps, and detection classification.
  • Emergency Evacuation Mobile App: Automatically broadcasts building alerts, provides interactive evacuation floor plans, and compiles real-time digital muster roll headcounts for fire wardens.

5. Analytics, Media Downloads & Retention

Manage your forensic recordings and monitor infrastructure health through integrated reporting tools:

Detection History & 1080p Downloads

Filter events by model class (Fire, Human, Car, Animal), camera, and date range. Download full-resolution MP4 clips with single-click actions.

Data Analytics & Gateway Telemetry

View incident heatmaps, false-alarm reduction trends, storage usage, and active Raspberry Pi Gateway health statistics.

Configurable Retention Policies: Set per-account video retention between 1 and 90 days. Automated background FFmpeg and storage workers securely purge expired media to ensure strict GDPR compliance.

6. Self-Service Billing, Security & Governance

  • Stripe Customer Portal: Upgrade camera quotas, update payment methods, and download VAT invoices self-service from your profile.
  • UK Data Sovereignty: Core AI engines and media storage are hosted exclusively on enterprise Vultr servers in Manchester, United Kingdom.
  • Enterprise Security (OWASP Top 10): Role-based access controls, immutable activity audit trails in Convex, and automated weekly vulnerability dependency scans.
  • 7-Day Account Deletion Grace Period: Self-service account termination includes a 7-day safety window to prevent accidental data erasure.
  • Autonomous AI Agent & OpenAPI Support: Documented OpenAPI schemas, agent.json, and /api/video-updates Server-Sent Events (SSE) for automated B2B integrations.

Need Setup Assistance?

Our UK technical support engineers can help you configure your RTSP cameras, deploy Pi Gateways, or set up emergency evacuation workflows.