Create a human interface device capable of sending OSC commands:

- Functions: Play/Pause, Forward/Reverse, Tap Tempo, Next/Prev, CUE#, etc...

- Output through ethernet, wifi, or usb.

- Could be a python script to capture keystrokes, or a smartphone app, or raspberry pi/arduino/esp8266

Create a server application:

- Functions: Load video files, output RTP

- Possibly Python wrapper for gstreamer

- Stream 640x480 mp4 with sub 50ms latency

- Switch video cues and change playback speed on the fly

Everything could be running on a single computer or distributed across a LAN. For optimal streaming latency probably will need dedicated network, either router or loopback interface.