hllshiro

同样需求，事实上开发量很小，希望能尽快集成！

So, how do you modify it when deploying with docker?

   可以通过自定义模型功能，你能够自己创建一个支持任何特性的模型。注意在创建时需要输入一个其他的名称，否则会选中已有的模型，然后在修改配置时修改为你期望的模型名称。

I also came to consult because it may not get the expected results when dragging. As in the example below, the two groups in the middle are locked, and after...

I need this too. The main purpose is to divide the process in two parts: Q&A guidance and result generation. Their prompts are the same, but the latter only outputs...

It seems that it can be implemented using node `HTTP Request`. The API interface will be provided after the agent is released.

> do you mean call another agent via HTTP request in workflow? it is smart, would be good if have this as a built-in feature. Yes, since dify doesn't provide...

I haven't looked at this problem for a long time, and tried many methods at that time without success. Now I have switched to [linkwarden](https://github.com/linkwarden/linkwarden) that have similar functions.

Something else, a widget developed by yourself, used in as chat bot in wechat. I've circumvented this issue in other ways at the moment. see [issue 8054](https://github.com/langgenius/dify/discussions/8054#discussioncomment-10585330)

是否有这方面的计划？