How to Add AI to the Security Cameras You Already Have

Organizations don't have to choose between doing nothing and replacing every camera. Adding video intelligence to the cameras you already own can turn a recording system into one that helps your team notice moments while they are happening.

This guide gives IT professionals and security integrators a practical framework for adding VOLT to an existing camera system while protecting current investments and keeping operations running. For structured help, see our school security system implementation plan, with templates and checklists.

Pre-Integration Assessment

A good integration starts with a clear picture of your current infrastructure and goals. The assessment sets deployment strategy, surfaces challenges early and defines what success looks like.

Begin by documenting all camera locations, models and capabilities. Record network details including bandwidth, switch configurations and storage.

Map your current monitoring workflows, response protocols and staff responsibilities so you know how alerts will fit in. Our AI security camera software budget preparation checklist can help with planning.

Current Infrastructure Checklist

• Camera Inventory: Models, locations, resolution and RTSP support for each camera
• Network Assessment: Bandwidth capacity (2-5 Mbps per camera for cloud deployment), switch configurations and data flow
• Storage Analysis: Current video storage, retention policies and capacity
• Monitoring Workflows: Existing security processes, response protocols and staff responsibilities
• Cameras: VOLT is built to run on the cameras you already have.

Schools can also use our VOLT for Schools brochure for education-specific considerations.

Deployment Architecture: Cloud vs. Edge

VOLT offers two deployment options for different network constraints and data requirements. Both keep your existing cameras.

VOLT's cloud option streams encrypted camera feeds to secure AWS infrastructure for real-time analysis. The on-premises edge option processes video locally on VOLT's purpose-built 1U server.

VOLT Cloud Deployment

• Network Requirements: Upload bandwidth depends on the number of camera streams and their configuration
• Processing Location: Secure AWS cloud infrastructure with enterprise-grade security protocols
• Management Interface: The VOLT Portal, accessible from any device
• Connectivity: A P2P connection from the edge device to VOLT's cloud, which may require port forwarding, static port mapping, or UPnP/NAT-PMP
• Updates: Software updates and security patches deployed automatically, without on-site visits

VOLT Edge Deployment

VOLT's edge option uses a purpose-built 1U server that runs event detection locally and installs in your server room. It gives you more data control and less dependence on the network.

Network Infrastructure Requirements

For either option, VOLT's edge device needs internet access and access to your cameras. If your camera network is separate and has no internet access, the device connects to both networks with minimum 1Gbps connections.

• Camera Connectivity: VOLT needs access to camera RTSP endpoints
• Firewall Configuration: Specific firewall exceptions let the edge device reach VOLT's core services
• Security Protocols: Allow the domains used for security patches and software updates

Deployment Timeline and Support

Timelines depend on site size and deployment type. VOLT's team supports camera configuration and portal setup.

Camera Configuration and Compatibility

VOLT works with existing IP cameras that support RTSP (Real Time Streaming Protocol). Proper configuration keeps streams stable and gives VOLT clear video to work with.

See our weapon detection with VOLT video for practical examples.

VOLT Camera Requirements:

• Streaming Protocol: RTSP
• Video Encoding: H.265 preferred, H.264 supported
• Resolution: 1920x1080p preferred, up to 2560x1440p supported
• Frame Rate: 10 FPS minimum, 20-30 FPS supported
• Bitrate: Under 3000 kb/s preferred for stable streaming

Resolution Guidelines

VOLT supports resolutions from 720p to 4K, with 1080p (1920x1080) ideal for most sites. Higher resolutions need more bandwidth and processing.

• Ideal: 1080p (1920x1080), 2K (2560x1440) and 4MP cameras
• Not Recommended: Below 720p or above 4K
• Not Supported: Above 8K (7680x4320)

Configuration Best Practices

• Encoding: H.265 preferred or H.264, with consistent bitrate under 3000 kb/s
• Stream Configuration: RTSP settings (usually under 'Stream Profile') matching VOLT's requirements
• Frame Rate: At least 10 FPS
• Resolution: 1920x1080p, up to 2560x1440p supported
• Network Stability: Stable connections for continuous analysis

Camera Positioning

Network Requirements

Requirements vary with the number of streams and your deployment option. The figures below assume camera streams follow VOLT's recommendations.

If your camera network can't reach the internet, the edge device connects to both networks with minimum 1Gbps connections. If cameras are on an internet-accessible network, one 1Gbps link is enough.

VOLT Cloud Deployment Bandwidth

The cloud option continuously streams encrypted camera feeds for processing. Estimated average upload bandwidth for the NUC:

VOLT On-Premises Edge Bandwidth

The on-prem device uses under 0.1 Mbps when idle. Data and video are uploaded only when a moment is flagged, when you watch live streams, or during historical playback. When a moment is flagged, expect an average spike of about 0.7 Mbps per camera stream involved:

Firewall Configuration

Add these exceptions so the VOLT edge device can reach VOLT's core services:

• Camera RTSP Endpoint: Port 554, required for VOLT's core services
• VOLT in AWS: Ports 50005, 443 (*.volt.ai, *.amazonaws.com)
• Security Updates: Ports 443, 80 (*.ubuntu.com, *.github.io, *.docker.com)
• Network Time Protocol: Port 123 (ntp.volt.ai)
• LaunchDarkly: TLS 1.2, Port 443 (https://*.launchdarkly.com)
• Sentry: Port 443 (sentry.io, *.sentry.io)
• Speedtest: Ports 443, 5060, 8080 (*.speedtest.net) to verify bandwidth
• Hardware Monitoring: Ports 10051, 10050 (monitoring.volt.ai)

Cloud Connectivity

The cloud option needs a P2P connection from the edge NUC to VOLT's cloud. Depending on your network, that may mean port forwarding/static port mapping or enabling UPnP, NAT-PMP, or PCP (NAT-PMP v2).

• Port Forwarding: Static port mapping for direct connectivity
• UPnP/NAT-PMP: Automatic port configuration (review security considerations)
• Tailscale VPN: Port 443 outbound from port 41641, for remote management if AWS is disrupted

Network Best Practices

• VLAN Segmentation: Keep camera traffic on dedicated segments routed to the VOLT edge device
• QoS: Prioritize VOLT video streams over general traffic
• Redundancy: Plan backup network paths for critical camera feeds
• Bandwidth Monitoring: Watch performance and catch issues early
• Security Controls: Firewall rules specific to VOLT's requirements

For privacy questions: VOLT doesn't share data or AI insights with law enforcement unless the customer requests it, and the customer owns the data. See also our article on AI video privacy considerations.

Implementation Phases

A phased rollout reduces risk and gives teams time to refine their processes.

Phase 1: Pilot

Start with 10-20 cameras covering a mix of areas: busy hallways or lobbies, restricted spaces and places with tricky lighting.

• Technical Validation: Confirm camera compatibility and network performance
• Process Testing: Walk through alert routing, response steps and staff onboarding
• Baselines: Record how incidents are handled today so you can compare later
• Feedback: Gather input from security staff, IT and end users

Phase 2: Zone-Based Expansion

Expand to more areas based on priorities and what you learned in the pilot. See our perimeter security with VOLT demonstration for building exteriors and parking areas.

Phase 3: Full Deployment

Complete deployment across remaining areas, confirm alert routing for every team, and set long-term maintenance and response procedures. Schools can see how VOLT fits the Standard Response Protocol in our guide on enhancing school safety through the SRP and VOLT.

• Complete Coverage: All planned cameras connected and managed in one portal
• Onboarding: Staff know how to receive, review and act on alerts
• Operating Procedures: Maintenance, monitoring and response protocols in place

Testing and Validation

Test both the technology and the workflow before full rollout, and again at regular intervals.

At Prescott High School, VOLT's system detected a student suffering an asthma attack in an empty hallway, enabling nurse response within 15 seconds

Reviewed by a person by default before your team is notified.

• Baselines: Document starting points for later comparison
• Scenario Testing: Run controlled tests of likely scenarios
• Integration Validation: Confirm communication with existing security systems and emergency protocols
• Alert Delivery: Confirm alerts reach the right people
• Documentation: Keep records of tests and results

Operational Integration and Staff Onboarding

Integration works when alerts fit the way your team already responds.

"This is the first time where I really have been able to be out ahead of things that are happening. I'm not just using my cameras for investigation, I'm using them for immediate action and response which is pretty special."

Adam Neely, Principal, Prescott High School

• System Operation: Portal navigation, alert management and response steps, with hands-on practice
• Integration Procedures: Coordination with existing security systems and emergency protocols
• Troubleshooting: Common issues and how to resolve them
• Continuous Learning: Ongoing updates as VOLT adds capabilities

Workflow Integration

Ongoing Review

• Alert Review: Look at what was flagged and how the team responded
• Operational Impact: Track how incidents are handled over time
• System Health: Monitor connectivity and maintenance needs
• Improvement: Adjust alert rules and procedures as needs change

See our VOLT for Corporate and VOLT for Real Estate brochures for other settings.

Getting More From the Cameras You Have

VOLT is built to run on the cameras you already have. It helps teams notice everyday moments such as crowds forming, theft, vandalism, and people in a space at times you set, and it detects handguns, long guns, and knives that are visible to the camera. Reviewed by a person by default before your team is notified.