A unified "input form" markup language should be added to all existing AI chats. Once the UI renderer of the AI chat recognizes such a markup language, it will render it into form elements such as buttons, options, and input boxes. In this way, when users reply, they don't have to write lengthy content completely. Instead, they can directly click on the options or buttons, or fill in the form to complete the input as a reply efficiently. This will significantly improve the reply efficiency of some chat interactions.

For example, when the MCP service requires users to supplement certain information, the LLM model will summarize it into the "input form" markup language, and users only need to quickly fill in this form to complete the reply.

Another example is that when custom instructions in a project or custom prompt files present some basic options for users to choose from, in the AI chat, these will be presented as clickable buttons instead of requiring users to manually input text for their replies.