Loading
Every tenth
counts.
WCA inspection · full stats · no account
Bring your solve history with you — import it in seconds.
0 turns
Why Cubeathon exists
A stopwatch tells you how long a solve took. It doesn't tell you why it took that long.
Getting faster isn't one skill — it runs in sequence every solve: reading the scramble, recognizing the case, executing the algorithm, and reviewing what happened once the timer stops. Most of that has nothing to do with the number on the clock.
Cubeathon is built around all of it. Timing is where a solve ends — it was never meant to be where practice begins.
1
Inspect
Read the scramble before you move.
2
Recognize
See the next case before your hands do.
3
Solve
Execute — the part the clock measures.
4
Review
What the session meant, not just its average.
5
Improve
Whether this month looks different from last.
Design principles
Timer First, Not Timer Only
The clock has to be correct before anything else matters — but it's the beginning of the model, not the whole of it.
Local by Default
Your solves live on your device unless you choose otherwise. Sync is a choice, not a requirement.
Built with Cubers
Priorities come from what practicing cubers actually run into, not a generic feature checklist.
Recognition Is Trainable
Recognition and inspection are practiced on purpose — not left as a side effect of doing enough solves.
Review Over a Single Time
One solve rarely matters. Whether the trend is moving is what the stats are actually for.
Serious tools.
Zero catches.
Every feature is free
Timer, stats, lessons — all free.
No ads, ever
No banners, no popups. Just you and the cube.
Works instantly
Open and solve. No signup, no download.
Your data stays yours
Saves locally. Export to JSON anytime.
Hold. Release. Fly.
Tap to stop.
R U R' U' F2 L D' B2 U2 R'
0.00
HOLD SPACE TO ARM
Shaped by conversations, not a spec sheet
A few themes kept appearing, across the discussions we followed on Reddit, Discord, and SpeedSolving: how to actually drill recognition instead of hoping it improves on its own. Whether inspection habits can be trained deliberately, not left to instinct. How to keep an algorithm set organized enough to review, instead of memorized once and forgotten. What a session should show you beyond a single average. Whether a month of practice actually moved the needle.
Already here
- Session review, not just a running average
- Algorithms organized by set, not one long list
- WCA-accurate stats you can actually trust
Actively being explored
- Dedicated recognition drills
- Guided inspection practice
- Longer-range progress tracking
Built in public beta
Public beta means the parts you can already use — the timer, the stats, the lessons — are built to last. It means the roadmap isn't finished yet, not that the timer isn't ready.
Your solves save to your device by default, not a server. Export everything to JSON whenever you want — no lock-in, no waiting on us.