ManiSkill is a leading open-source platform focusing on robotic simulation, unlimited robot data generation, and the generalization of robotic AI. Led by HillBot.ai, the platform supports rapid training of robots through state and/or visual inputs, achieving a 10-100x increase in visual data collection speed compared to other platforms like ManiSkill/SAPIEN. It supports parallel simulation and rendering of RGB-D images on GPUs, with speeds exceeding 30,000+ FPS. ManiSkill offers over 40 skills/tasks and more than 2,000 pre-built objects, along with millions of frames of demonstrations and dense reward functions, allowing users to focus solely on algorithm development without gathering assets or designing tasks themselves. Additionally, it supports simultaneous simulation of different objects and joints in each parallel environment, reducing the time for training generalized robotic strategies/AI from days to minutes. ManiSkill is easy to use, can be installed via pip, and provides a simple, flexible GUI along with extensive documentation for all functionalities.