Integration of DeepSeek V3 Model in RemNote

As a frequent RemNote user, I've been grappling with a frustration that I'm sure many others share: the alarming rate at which AI tokens are burned through. It's like having a sports car with a tiny gas tank - you have this amazing tool, but you can barely use it because it's constantly running on empty. By mid-month, I'm counting tokens like I'm counting pennies before payday, and this seriously hampers the potential of a tool that should be helping us study and work better. Integrating DeepSeek V3 could be that game-changing shift we all need - a more efficient model that would let us actually harness AI without constantly watching the token counter like a hawk