Quiz · 7 questions
🔌 Everything Is an Adapter
One narrow seam, many swappable engines and brains
Level 0Unlit
0 XP0/48 lessons0/13 achievements
0/100 XP to next level100 XP to go0% complete
Quiz
01Firekeeper puts its abstraction boundary in a few narrow places rather than making everything provider-neutral. Why is 'narrow seam' better than 'abstract everything'?
02The STT interface returns both text and a language. Why include the language in the return value instead of assuming it?
03The cleanup provider returns only the cleaned text, and that text is measured before insertion. Why must the insertion layer never receive arbitrary text from the cleanup brain?
04Connection modes (This Mac / Tailscale / Local Mini) are a settings choice, not a hidden fallback chain. What's the risk of making them an automatic silent fallback instead?
05Firekeeper vendors some family-kit files as generated copies (KitBrainControl.swift, the Python utility worker) instead of linking them. What makes that safe?
06Command mode needs a brain. Where does Firekeeper get it, and what happens if it cannot?
07Since 2026-09-21 dictation's Pippa utility lane runs in an app-owned worker on office. How does a MacBook reach it?
Comments 0
🔔 Reply notifications (sign in)Sign in — Please sign in to comment.
No comments yet — be the first.