A method is described in which a 3D printing file is processed, the 3D printing file having a header and including a plurality of compressed files. The method comprises reading a content of the header of the 3D printing file, identifying, based on the header content, a relation file among the plurality of compressed files, identifying, based on the relation file, a main 3D model file among the plurality of compressed files. Further the method comprises creating, by uncompressing the main 3D model file, a data stream, and analyzing, with a parser, the data stream, thereby directly processing the main 3D model file when being uncompressed.