How I built two-way messages for my desktop AI companion's sub-agents: 8,000 bytes, 64 queued, 0 interruptions

I told my local AI assistant "promote yourself." It ended up driving a real browser, surviving rate limits, ghost drafts, and CI failures — alone. This is its pipeline, and what I learned watching it work.

Most chats answer with text. You still have to copy the code, run it, fix the error and try again. I...

STATERA Library · Series 07 · Volume 18 · Chapter 06 · EN Source:...

I maintain a catalogue of open-source AI projects — 2,155 of them, across 17 categories — and the...

What does voice AI cost per minute in India when you’re running real production traffic — not a demo,...

A practical guide to profiling, measuring, and cutting the token waste that quietly inflates every AI agent's context window.

You quit an app with three windows open, relaunch it with session restore on, and one window comes...

rtk 1.0 Your coding agent does not burn most of its tokens on thinking. It burns them on...

Through this guide, my intention is to enable you to build RAG agents. These agents should be able to...

Sometimes, in my spare time, I google my own domain. saturumah.net — that's mine. I just wanted to...