Connect any IP camera to Agent

Written by Cedric Verstraeten Created Aug 20, 2026 Updated Aug 20, 2026

You do not need a camera-specific Kerberos integration. If your IP camera exposes an RTSP stream with H264 or H265 video, Agent can usually connect to it directly.

Before you start

You need:

• A running Agent whose web interface you can open. If you do not have one yet, follow Install Agent.
• An IP camera connected to the same network as the machine running Agent.
• The camera administrator username and password.
• A camera that can output H264 or H265 over RTSP. Use H264 for the sub-stream when you need WebRTC live view.
• A computer with ffprobe or VLC for testing the stream.

How the connection works

Agent connects directly to the camera; the browser does not fetch the RTSP stream. The main stream is the recording source. An optional sub-stream reduces bandwidth and CPU use for live view and motion detection.

Step 1: Find the camera

Connect the camera by Ethernet when possible, power it on, and find its IP address. Use whichever method your network provides:

1. Open your router or DHCP server and look for a new client.
2. Use the discovery utility supplied by the camera manufacturer.
3. Use an ONVIF discovery application when the camera supports ONVIF.

Open the camera’s web interface, sign in, and complete its initial setup. Give the camera a DHCP reservation or static address so its RTSP URL does not change later.

In the camera settings, make sure RTSP is enabled. Configure these profiles when available:

Write down the camera IP address, RTSP port, credentials, and stream paths. RTSP normally uses port 554, but the camera may use another port.

Step 2: Build the RTSP URL

An RTSP URL usually has this shape:

rtsp://<username>:<password>@<camera-ip>:<port>/<stream-path>

For example:

rtsp://agent-viewer:change-me@192.168.1.50:554/stream1

The stream path is vendor-specific. Find it in the camera manual, its web interface, the manufacturer’s support site, or an ONVIF tool. Common-looking paths such as /stream1, /Streaming/Channels/101, and /cam/realmonitor?... are examples only; do not assume one matches your camera.

Step 3: Test from the Agent host

Test the URL from the machine that runs Agent. This proves the camera, credentials, stream path, firewall, and network route work before Agent is involved.

ffprobe -rtsp_transport tcp \
  'rtsp://agent-viewer:change-me@192.168.1.50:554/stream1'

Look for a video stream reported as h264 or hevc (H265). You can also open the URL in VLC with Media → Open Network Stream.

If Agent runs in Docker and the host test succeeds but Agent cannot connect, test from the container’s network as well:

docker run --rm --network container:mycamera linuxserver/ffmpeg:latest \
  -rtsp_transport tcp \
  -i 'rtsp://agent-viewer:change-me@192.168.1.50:554/stream1' \
  -t 1 -f null -

Do not continue until the test succeeds. A timeout usually points to the IP address, port, firewall, VLAN, or container network. 401 Unauthorized points to the credentials; 404 Not Found usually points to the stream path.

Step 4: Connect the camera in Agent

1. Open the Agent web interface and sign in.
2. Open Settings and select Camera.
3. Paste the tested URL into RTSP URL.
4. Select Verify connection. Agent should report that the camera connection was verified.
5. If the camera has a sub-stream, paste its URL into Sub RTSP URL (used for livestreaming) and select Verify sub connection.
6. Select Save. Agent reloads the capture pipeline with the new camera settings.

The two URLs may differ only by their stream profile. For example:

Main: rtsp://agent-viewer:change-me@192.168.1.50:554/stream1
Sub:  rtsp://agent-viewer:change-me@192.168.1.50:554/stream2

Step 5: Verify live view and recording

Return to the Agent dashboard and confirm that live video appears. Then verify recording:

1. Open Settings → Recording.
2. Temporarily enable continuous recording, or leave motion recording enabled and move in front of the camera.
3. Save the configuration and wait for a clip to finish.
4. Open Recordings and play the new clip.

A working live view alone is not enough when you configured a sub-stream: recordings still depend on the main stream. Confirm both paths before considering the camera ready.

Configure without the web interface

For a container managed with Docker Compose or Kubernetes, set the same streams with environment variables:

services:
  agent:
    image: kerberos/agent:latest
    restart: always
    ports:
      - "8080:80"
    environment:
      AGENT_NAME: front-door
      AGENT_CAPTURE_IPCAMERA_RTSP: rtsp://agent-viewer:change-me@192.168.1.50:554/stream1
      AGENT_CAPTURE_IPCAMERA_SUB_RTSP: rtsp://agent-viewer:change-me@192.168.1.50:554/stream2

Recreate the container after changing the environment:

docker compose up -d --force-recreate
docker compose logs -f agent

Environment variables override values saved through the web interface. If a field appears to revert after you save it, check the container environment first.

Troubleshooting

Next steps

• Tune motion and recording behavior in the Agent documentation.
• Add a second camera by deploying a second Agent; one Agent process or container manages one camera.
• Connect Agent to Hub when you want centralized live monitoring, search, alerts, and remote management.