Labs runs 100+ models for image, video, music, voice and lipsync from a single interface, on one credit balance. References, masks, segmentation, upscaling and prompt generation are built in, so a result you like can be reworked without exporting it to another service.