Running RT Server ¶

RT Server can be run on both Windows and Linux platforms in headless mode, providing production-ready deployment options with REST API and Web Panel interfaces.

Windows ¶

Installation ¶

RT Server can be installed on Windows using the official installer from rt.cvedia.com.

The Windows installer automatically:

Installs RT Server (rtservice.exe) as a Windows service
Configures the service for automatic startup on system boot
Sets up the service to run in the background without user interaction

After installation, the RT Server service starts automatically and will be accessible via Web Panel at http://localhost:3546.

Command-Line Options ¶

The Windows installation consists of two components: the Downloader that fetches the installer files, and the Setup that performs the actual installation.

RT Server Downloader ¶

The downloader (CVEDIA-RT-Server-Downloader-<version>.exe) downloads and launches the RT Server installer for its bundled version. It supports the following parameters:

Parameter	Values	Default	Description
`/DOWNLOADONLY`	`true`, `false`	`false`	Download files without running the installer
`/VCREDIST`	`true`, `false`	`true`	Download VC++ Redistributables
`/IPBINDING`	IP address	(none)	IP binding to pass to RT Server setup
`/SILENT`	(flag)	(none)	Run in silent mode
`/VERYSILENT`	(flag)	(none)	Run in very silent mode (no UI)

RT Server Setup ¶

The installer (CVEDIA-RT-Server-Setup-<version>.exe) supports:

Parameter	Description
`/IPBINDING=<IP>`	Bind server to specific IP address
`/SILENT`	Run installer in silent mode
`/VERYSILENT`	Run installer in very silent mode (no prompts)

IP Binding Options:

127.0.0.1 - Local access only (default)
0.0.0.0 - Accessible from all network interfaces
<specific IP> - Bind to a specific network interface

Usage Examples ¶

:: Download and install (GUI mode)
CVEDIA-RT-Server-Downloader-2026.1.0.exe

:: Silent installation with network-accessible binding
:: Use "start /wait" to block until installation completes
start /wait "" "CVEDIA-RT-Server-Downloader-2026.1.0.exe" /VERYSILENT /IPBINDING=0.0.0.0

:: Download only (no installation)
CVEDIA-RT-Server-Downloader-2026.1.0.exe /DOWNLOADONLY=true

Linux ¶

RT Server on Linux can be deployed in three ways:

AI ApplianceNative SetupDocker

The recommended way to deploy RT Server on Linux is using the CVEDIA AI Appliance ISO image.

The custom ISO includes the full RT Server environment, all required dependencies, and all necessary system configurations. This allows you to perform a clean, ready-to-use installation with minimal setup.

Benefits:

Better performance (no container overhead)
Direct system integration with systemd
Easier log management and monitoring
Direct hardware access
Simpler networking configuration

See the AI Appliance Installation Guide for detailed instructions.

For advanced users who need to install RT Server on an existing Linux system.

Advanced Installation

This installation method is considered advanced and is not recommended for typical deployments. Proceed only if your use case requires a nonstandard setup.

Installation ¶

Install RT Server using the convenience script:

curl -sqko - http://get.cvedia.com/rt-server | sudo bash

Environment Variables ¶

The installation script supports the following environment variables:

Variable	Description
`CVEDIA_WEB_SERVER_IP`	Override the RT Server binding IP (e.g., `0.0.0.0` for all interfaces) - Default is `127.0.0.1`
`CVEDIA_ACCEPT_EULA`	Set to `1` to automatically accept the CVEDIA's EULA

Example with environment variables:

curl -sqko - http://get.cvedia.com/rt-server | sudo CVEDIA_WEB_SERVER_IP=0.0.0.0 CVEDIA_ACCEPT_EULA=1 bash

Running RT Server ¶

Method 1: Using systemd service (recommended for production)

# Start the RT Server service
sudo systemctl start cvedia-rt

# Enable automatic startup on boot
sudo systemctl enable cvedia-rt

# Check service status
sudo systemctl status cvedia-rt

# Stop the service
sudo systemctl stop cvedia-rt

Method 2: Running directly with rtservice binary

# Run RT Server directly
rtservice

# Run with custom configuration
rtservice --webserver-port 8080 --solutions /custom/path/solutions

# Run in background (using nohup)
nohup rtservice > /var/log/cvedia-rt.log 2>&1 &

File Locations ¶

The native installation places files at:

Installation directory: /opt/cvedia-rt/
Configuration: /opt/cvedia-rt/rtconfig.json
Solutions: /opt/cvedia-rt/solutions/
Instances: /opt/cvedia-rt/instances/

RT Server can run within Docker containers using the run.sh script in service mode.

Running in Background (Daemon Mode)¶

# Run RT Server in background
./run.sh --daemon

# Run with specific API port
./run.sh --daemon --api_port 3546

# Run without exposing any ports
./run.sh --daemon --no_expose

Service Options ¶

--api_port <port>: Port to expose API, default: 8080; Use 0 to disable exposing.

--rtsp_port <port>: Port to expose RTSP stream, default: 8554; Use 0 to disable exposing.

--mqtt_port <port>: Port to expose MQTT service, default: 1883; Use 0 to disable exposing; Ignored when --disable_mqtt is set.

-E, --no_expose: Do not expose any service or port; This overrides all other service / port expose options.

--disable_mqtt: Disable internal MQTT service

--use_image_rest_ep: Enables synchronous image rest ingestion endpoint. Requires internal MQTT service.

--intent <value>: Docker intent, possible options are:

run: Run the container (default)

create: Create the container but do not start it

Key Service Parameters ¶

--daemon: Run container in background as service
--api_port 3546: Expose REST API on port 3546 (default: 8080)
--no_expose: Don't expose any ports (security mode)
--webserver: Enable web admin panel (enabled by default)

Managing Docker Service ¶

# View running containers
docker ps

# View logs from background RT Server
docker logs cvedia-rt_<container_name>

# Stop background RT Server
docker stop cvedia-rt_<container_name>

# Remove stopped container
docker rm cvedia-rt_<container_name>

Configuration ¶

RT Server behavior can be configured through the global configuration file rtconfig.json. See Global Config for details.

Some settings can also be overridden using the rtcmd command-line tool. See rtcmd Configuration Management for details.

Web Server Configuration ¶

The embedded web server can be configured in rtconfig.json:

{
  "webserver": {
    "enabled": true,
    "port": 3546,
    "host": "0.0.0.0"
  }
}

Auto-Discovery ¶

RT Server includes auto-discovery features that can be configured:

{
  "auto_discovery": {
    "enabled": true,
    "port": 12349
  }
}

rtservice - RT Server Process ¶

The RT Server process is available as rtservice.exe on Windows and rtservice on Linux. It can run as a foreground process or as a system service (Windows service or Linux systemd).

Installation locations:

Platform	Executable Path
Windows	`C:\Program Files\CVEDIA\CVEDIA-RT\rtservice.exe`
Linux	`/opt/cvedia-rt/rtservice`

Use rtservice --help for the most up-to-date options:

rtservice [OPTIONS]

OPTIONS:
  -h,     --help              Print this help message and exit
  -s,     --solutions TEXT:DIR
                              Path to the AI solutions folder
                              Windows: C:/ProgramData/CVEDIA/CVEDIA-RT/files/solutions
                              Linux:   /opt/cvedia-rt/solutions
  -i,     --instances TEXT:DIR
                              Path to the instances folder
                              Windows: C:/ProgramData/CVEDIA/CVEDIA-RT/files/instances
                              Linux:   /opt/cvedia-rt/instances
  -d,     --daemon            Run as a system service
          --auto-discovery, --no-auto-discovery{false} [1]
                              RT auto discovery
          --version           Print version

Logging:
          --log-level TEXT:{verbose,debug,info,warning,error,fatal} [info]  (Env:RT_LOG_LEVEL)
                              Set log level
          --log-max-size INT [52428800]
                              Log file size at which to rotate the log
          --log-max-files INT [3]
                              Number of rotated logs to keep
          --log-file TEXT [cvediart.log]  (Env:RT_LOG_FILE)
                              Set log file
          --log-console       Enable console logging

Webserver:
          --webserver, --no-webserver{false} [1]
                              Enable webserver for REST API, Admin panel and HLS
          --webserver-host TEXT:IPV4 [0.0.0.0]
                              Host to listen on
          --webserver-port INT:INT in [1 - 65535] [3546]
                              Port to listen on

Performance:
          --pin-on-core       Assign a random but fixed CPU core to each instance
          --opencv-num-threads INT:INT in [-1 - 64] [0]
                              Number of threads to use for OpenCV

Plugins:
          --plugin-list       List all loaded plugins

Key Options ¶

--daemon: Run as system service (Windows service or Linux systemd)
--webserver-port: Change default Web Panel port (default: 3546)
--webserver-host: Set binding interface (default: all interfaces)
--log-console: Enable console logging for debugging

Production Considerations ¶

Security ¶

RT Server exposes web services by default on port 3546
Consider firewall rules to restrict access
Use --webserver-host to bind to specific interfaces
Disable auto-discovery in production: --no-auto-discovery

Performance ¶

Use --pin-on-core to assign CPU cores to instances
Configure --opencv-num-threads for optimal OpenCV performance

Monitoring ¶

Enable console logging: --log-console
Configure log rotation: --log-max-size and --log-max-files
Use REST API health endpoints for monitoring
Access Web Panel at http://server-ip:3546

Next Steps ¶

Web Panel - Learn about the browser-based interface
REST API Integration - Programmatic control and monitoring