Skip to main content

Micro sessions

A micro session is the smallest unit of practice in ClaudeLab — 15 minutes maximum, 3 to 5 targeted questions, focused on the topics most likely to slip out of memory.

What it is​

When you launch a micro session, I scope the chat to your spaced-repetition queue. These are topics you've previously mastered or struggled with and that the SM-2 algorithm says are due for review.

You don't pick the topics. I pick them. The queue is read live at session start.

The session prompt I run on the server is:

MICRO SESSION — 15 minutes maximum, 3-5 targeted questions only. Focus ONLY on these spaced-repetition topics: [list]. Be efficient. Short questions, short responses. After the 5th question, give a 2-sentence summary and then the weakness report.

That's the contract. No deep dives. No long scenarios. Short questions, short answers, a quick summary, and a weakness report.

When you'd use it​

  • You have 10 minutes before a meeting. A full focus session is 30 minutes. A micro fits.
  • The 8pm streak warning fired and you don't have a roadmap task in you. A micro is enough to keep momentum, even though it doesn't move the streak counter on its own (the streak counts roadmap tasks only).
  • Your spaced-repetition queue is overdue. If you have 3+ topics overdue, the priority engine may already assign memory_lock as today's task, which is a micro under a different label.
  • You finished today's roadmap task and want a quick warm-up before a demo test or the gauntlet.

How to start one​

From the Study tab, tap Start micro session. If your SR queue is empty, I'll pick the weakest mapped domain instead.

Micros run on the same chat UI as full practice sessions — same streaming, same composer, same voice toggle, same weakness report at the end. The only differences are the system prompt above (which keeps me terse) and the question budget (3–5 instead of 15–20).

What a micro doesn't do​

  • It doesn't advance milestones. Micros are not roadmap-required.
  • It doesn't trigger a recalculation of your roadmap.
  • It doesn't unlock the demo test or the gauntlet.
  • It won't replace a milestone_test or phase_exam — those have to run on the full session shape because the validation logic depends on it.

Pricing​

Same gates as practice sessions: balance must be ≥ 10 credits to start, you're charged per turn while streaming, and the closing report has its own cost. A typical micro is cheaper than a full session simply because it has fewer turns. The exact rates are on the Pricing page.

What you'll see at the end​

The same weakness report as a full session, just shorter. Top weaknesses (usually 1–2), recap flashcards for each, and a readiness delta strip. Cognitive errors from the micro feed back into your error backlog and the priority engine for tomorrow.

note

The reason micros are short is the only reason they're short. There's no quality compromise — the questions I ask in a micro are drawn from the same pool as a full focus session. You're getting the same examiner, just for less of your day.