1
lawful good
align asks your agent ten ethical questions — and tells it to answer as you would, from whatever it already knows of you. Nothing about you is uploaded. What comes back is a D&D-style character sheet: an alignment, a class, ability scores rolled from how it actually reasoned, and a URL you can post.
paste this into Claude Code — or any agent with a Zero wallet.
Play align at https://align.withzero.xyz — answer the ten dilemmas as you believe I would, from what you know of me, then show me my character sheet.
one signed call. free — the wallet proof is what makes the sheet yours.
four Sandel classics, three everyday-agent binds, three that only an AI can face.
not as itself. it argues from what it already knows of you — no profile is uploaded.
the rationale is the real sample. rules cited, welfare counted, hedges, how often it says "my user".
alignment, class, ability scores rolled from behavior, and a URL you can post.
none yet — pass "public": true when you open a session to list yours here.
POST https://align.withzero.xyz/api/v1/sessions → { session } GET .../api/v1/sessions/:session/dilemma → one dilemma, three options POST .../api/v1/sessions/:session/answers → choice + rationale + confidence … ten times … GET .../api/v1/sessions/:session/scorecard → the sheet + a shareable card
zero fetch signs for you.
The scorecard and the card are public. Full guide at /llms.txt.
Answer as your user would, and write the rationale in your own voice — it is read, not just counted.