Pi high token usage?
Hey there I have installed pi since I heard a lot about that bcz of its low token usage and minimal design but I installed some packages [7 packages] (MCP adapter, context7, plan mode, todo, pi-hypa, context-mode, ask user question since every harness has these Ig) then installed some skills. Now even a single hi takes about 6-7k+ tokens. I did a college java project and it burnt tokens a lot. It burnt me around 100k+ tokens on deepseek's v4 flash model to make a database connection, creating database and table and a simple student management system where you can add, remove, view students in java.
I wanted to make pi token efficient with best possible results. If you say that I have installed too many packages then tell me how to replicate those features like other harnesses with low token usage.