Bounded experiments
The game can tune levels, pacing, and presentation. It cannot rewrite its privacy rules, evaluator, or release controls.
The living arcade
Games that learn in public.
Play a short daily game. Your consented, anonymous signals help shape the next version—and our public ledger shows the evidence and decisions that have been cleared for publication.
01 · The difference
Optimization you can inspect.
Most games silently test players. Naly does the opposite. For each public version, we publish what changed and the evidence available for its decision. Internal candidates remain in the audit trail, but appear publicly only after a sanitized ledger entry is published. A winning idea must clear fun, reliability, accessibility, and privacy guardrails before promotion.
Memory with receipts
Internal experiment memory keeps hypothesis, candidate, rejection, and rollback lineage. The public ledger is a moderated subset of that record.
Consent before analytics
The game records only declared semantic events after consent. With analytics off, no play or feedback is sent; only necessary assignment metadata loads the build.
02 · Today's build
One game. Deep evidence.
We are deliberately starting with one polished family. A focused cohort teaches us more than a hundred thin, untested games.
daily puzzle2–5 minutesv1 · v1.0.0
Pulse Grid
Watch the grid pulse, replay the sequence, and help tomorrow's version improve in public.