GePi should be able to serve incoming requests. There is no necessity that the UI already works completely as long as the result tables can be downloaded. The download should contain the statistics sheet that we created with the Jupyter Notebook for the last runs done with traditional UIMA pipelines + result writer.