Add this as global config: https://www.reddit.com/r/ChatGPT/comments/133mc37/experimental_how_to_compress_your_prompts_so_you/

In chat messages list we would to load only N messages to prevent useless historic data.

As title says

V0.1

• Make OpenAI flow working
• Prepare client packages
• Tests!
• Do some code cleaning

V0.2

• Make a light version

V0.3

• Allow custom CSS themes

V0.4

• Allow custom CSS themes

V0.5

• You can load your own LLM locally or load it from remote (as LLM profiles)

In order to enhance this, we can add a Stemming process to optimize messages

It would be great if we provide an autocomplete on input and fill everything with [TAB]

It will supports custom commands and prompts

You can toggle the option to reduce API consumption

We need an option to set a default key

There's something wrong with the API stream and nativephp, causing no working more the app. Need to investigate.

We need to bypass/ignore the AI response when it returns an image (or any non-text).

Just to prevent malformed apis

We need tests Boyz

Prefer Pest

We need to pack all custom phrases into localizable, then we need to define guidelines of translations