VideoAdGuard by 一只小包子

Based on the large language model, automatically identify the implant/oral broadcast advertisement in the Bilibili video. When the video enters the oral broadcast advertisement that is not related to the content, the advertisement can be automatically skipped. The API key of the large model is required, and all information is saved locally, so there is no need to worry about information leakage. The code is open-sourced in https://github.com/Warma10032/VideoAdGuard. If you have any questions, please go to Github to discuss. If you want to experience it for free, you can use the glm-4.5-flash free model of Zhipu AI, which works well.