Typing Mind

We can have option under each results to chose which results to reinject in discussion with LLM (to save tokens) By default all previous results are send back to LLM, but the more we interact, we don’t want somes to pollute the discussion… I know we already have option to send n previous messages, this one is more granular

I would like to be able to split chats to separate the context. It is common for me to find my conversation has drifted and the responses sometimes feel (or clearly are) wrongly influenced by stuff earlier in the context window. Being able to split them would avoid having to fork and/or delete a bunch off interactions.

OpenAI can respond with multiple responses after sending one prompt. Example would be - I have a large large dialog, lots of tokens in use, and I am looking for a specific code solutions. But if gpt gives me poor code - I need to rerun it again and spend all that money on the prompt again. Or I can use something like n=3 and have 3 responses to try for the price of one prompt. Alternative use - if I need 10 ads written, Id rather use n=10 instead of asking for 10 ads in one response, because every next one is influenced by the previous one.

I'm requesting a feature enabling automatic repetition of certain prompts, specifically the "continue." command, with the ability to loop this up to 20 times. An infinite loop would be too risky, but a cap at 20 provides a good balance. Additionally, for those cautious about this functionality, a setting to hard lock the limit, or disable the looping feature entirely, would be ideal. This would greatly streamline tasks requiring continuous follow-ups and minimize the need for constant manual prompting.

Rules to automatically add conversations with tags to a certain folder, or add tags to conversations when added to a specific folder.

Due to a bug, the code was not displayed correctly. After the hotfix, it is shown as plain text, which works quite well, but is harder to read. Therefore, it would be nice if syntax highlighting also works for Razor and of course Blazor.