How effective is the "Brain" feature in ManyGPT for daily workflows?

ManyGPT introduces the "Brain" feature to help centralize context and streamline AI-driven tasks, but I'm curious about how well it holds up under heavy daily usage.

For early stage products adding memory or custom context modules, the main challenge is usually balancing accuracy without overwhelming the user with setup complexity or getting irrelevant outputs over time.

For anyone who has been using ManyGPT:

How well does the Brain feature handle long-term context retention in your actual setup? Is it smooth enough for complex projects, or are there areas where it still needs tweaking?