[2h] Add Manual Camera Configuration Input

Currently, camera selection is limited to an automatically discovered list. This approach fails when cameras are located on a different subnet, as they are not detected, limiting functionality for users with non-standard network configurations.

Look for the camera discovery and selection logic in the frontend codebase, likely in a component or service handling device connections. The task is to add a manual input form for camera configuration (e.g., IP address, port). Check existing UI patterns for form inputs and network configuration. Verify the change by testing that a manually entered camera can connect.

Written by the indexing model from the issue text.