Boards

Feature suggestions

Problem feedback

Powered by Canny

Feature suggestions

Please describe only one function point per article and try to be as brief as possible. Keep an eye out for similar suggestions, and comment or vote if you already have them.

Various “URL - model name” bindings are pre-saved, and quickly switched during use

I have many models on multiple servers, and when switching calls on ChatBox, I have to re-enter the URL address every time. I hope it can be like a “custom model name”

complete

There are a few options to increase the number of messages in the long run. For example, add an option of 0-1000

Now, with the exception of 0-20, that is, unlimited. For long-form novels, a message would cost 160,000 tokens yesterday, but if you open it to 20, you'll quickly lose your memory. There's really too little context for 0-1000, even 0-200. 20 is really too little context. It can't meet diverse needs and multiple scenarios.

complete

The thought process is closed by default

When using the inference model, every time you click on the dialog box, the thought process unfolds all by default, which greatly affects finding contextual content. Collapsing one by one is very troublesome. Can you add a default Collapse switch or set a one-click Collapse or Hide Thinking Process switch

complete

Currently, there are too few models that can be connected to the mobile terminal, please sync to the same as the PC version as soon as possible

complete

Can the networking function always be on or off? Every time you connect, you have to click extra, and can it support Google search

complete

Powered by Canny