• AI model optimization: Improved speed, and will automatically decide how to use inference based on the task type;
  • Xiaohongshu analysis and optimization: improved the success rate of obtaining content requiring a login page;
  • Zhihu Q&A and column analysis optimization: faster analysis into articles is supported;
  • Minority analysis optimization: more stable retrieval of web content;
  • The reader's file preview function has been adjusted: Changed to support only PDF type files. Other temporary files will only support download, and direct preview will not be supported;
  • Fixed an issue where some content or images were lost when saving chat history on WeChat;
  • Fixed an issue where some email sender names were garbled.