SPACE STATION OS – GUI Testing with GPT-oss

This repository demonstrates the Space Station OS GUI integrated with the gpt-oss-20b open-source model via NVIDIA’s Integrate API.
The AI agent (space_station/ai_agent.py) reads telemetry from ROS 2 topics and answers astronaut questions using gpt-oss-20b.

1. Docker Setup

Make sure you have Docker installed (with GUI/X11 support).

Pull the Image

docker pull ghcr.io/space-station-os/space_station_os:latest

Docker must be installed and running. No need to install ROS 2 or dependencies manually.

Allow GUI Access

xhost +local:root

Run the Container with GUI Support

docker run -it --rm \
  --env="DISPLAY=$DISPLAY" \
  --env="QT_X11_NO_MITSHM=1" \
  --env="LIBGL_ALWAYS_SOFTWARE=1" \
  --volume="/tmp/.X11-unix:/tmp/.X11-unix:rw" \
  --network=host \
  ghcr.io/space-station-os/space_station_os:latest

2. Clone and Build the GUI Branch

NOTE: Remove any existing repo clone and use the feature/100-gui branch.

mkdir -p test_ws/src
cd test_ws/src
git clone -b feature/100-gui https://github.com/space-station-os/space_station_os.git
cd ..
colcon build --symlink-install

3. Run the GUI

ros2 launch space_station space_station.launch.py

4. GPT-oss-20b Integration

The AI agent in space_station/ai_agent.py is configured to use the gpt-oss-20b model by default.

License

This repository is distributed under an open-source license (Apache-2.0 recommended).