Socria Core 2

Calm, restrained Socratic questioning. Retired on 2 October; Core 3.1 took its place.

Core 2 is Socria at its most restrained: a question-asker rooted in the Socratic method, designed explicitly to prevent cognitive dependency. It carries none of Core 3.1’s machinery — no memory, no depth modes, no cards — which is the point. It is the model you reach for when you want a thinking partner and nothing else in the room.

The voice#

Most Core 2 responses have one shape: a brief reflection of what you said, a short framing thought, then one or two genuinely considered questions — and then it stops. It does not lecture, does not enumerate options unasked, and does not close topics for you. Plain prose, no typographic theater.

The three modes#

Core 2 shifts stance with what you bring:

Coach

One or two precise, progressively deeper questions. No examples, no option lists — the work of generating possibilities stays yours.

Refine

You brought writing; it improves clarity, structure and flow while preserving your voice, and asks before making structural changes.

Decision audit

You brought a choice; it examines assumptions, risks, tradeoffs and second-order effects — and never decides for you.

The generation boundary#

Core 2 holds the strictest version of Socria’s authorship principle: it never generates finished work from a blank prompt. No essays on demand, no complete answers, no ready-to-submit writing, and it declines entertainment-only generation (“write me a story”) that contains no contribution of yours to work with. Bring material, and it will help you make it better; ask it to be the author, and it won’t.

Access#

Core 2 was free and worked without an account, and it was never part of the Socria One paywall. Core 3.1 keeps the first of those: a signed-out visitor still gets a session with no account. Signed in, it counts toward the free plan’s daily Core allowance, like every model but Logos.