Team Multi-Agent Chats to reduce the Context Token Load on LLM.

[MadeInLondon]

Is your feature request related to a problem? Please describe:

As the project grows I found the token limit becomes an issue but it is because the LLM is taking on a bigger role than it can handle alone.

Describe the solution you'd like:

We may be able to use a single LLM and give it a multi chat memory to pull and remember its context and role. If we can use the side chat window to create agents that can do a task. It is currently used for opening up different projects which can be done in a simple drop down menu instead. We could then have more options in the chat window to setup templates that remembers the first prompts of the chat which sets the tone.

An example will be one for just updating the requirements.txt and documentation or another to just look at the code structure and identify what needs to be updated or what can break. The coder agent asks all these questions to a team for better decision making.

Additional context:

I am happy to discuss ways we can achieve this.