Connect matching nodes across the grid.
Solve the built-in levels.
Design your own grids.
Design custom paths.
Manage your grids and track your solutions.
Publish grids and test community creations.
Explore the mathematics behind grids.
Pick one of the revealed starter grids. Random grids now live in the Level Editor.
A mathematical lens on size, structure, solvability, and why local tile rules transform global difficulty.
Core idea
Size alone does not determine difficulty. A larger board can be easier if it creates more routing slack, more alternate corridors, or fewer forced collisions. A smaller board can be brutal when a handful of local choices controls the entire network.
In other words, hardness often comes from structure, not scale: bottlenecks, endpoint ordering, symmetry, local rigidity, and how much freedom each tile preserves for the rest of the puzzle.
Before asking for the best route, ask whether any valid route exists at all. Boundary placement, disconnected regions, narrow bottlenecks, and forced pairings can make a puzzle impossible before the search truly begins.
Which local constraints would let you prove impossibility before doing a full search? Which bottlenecks or disconnected regions behave like instant certificates of failure?
Some grids collapse to a single rigid routing. Others admit many valid completions. Counting solutions tells us how constrained a puzzle really is, and whether the challenge comes from logic, exploration, or ambiguity.
What changes turn a rigid grid into one with many valid completions? Could you estimate the number of solutions from symmetry, branch points, or corridor width before enumerating every route?
Among all valid routings, which one is the cleanest? Simplicity might mean fewer turns, fewer detours, more symmetry, or more natural pairings. The simplest solution is not always the one that is easiest to discover.
How should simplicity be measured: fewest turns, shortest total path length, strongest symmetry, or the most balanced pairing pattern? Different metrics may prefer very different “best” solutions.
The key issue is flexibility. A two-path tile makes each placement a strong commitment, and those commitments propagate across the board. A three-path tile changes the global behavior because it introduces more slack, more rerouting options, and fewer brittle dead ends.
Where does extra connectivity stop being “more complexity” and start becoming “more freedom”? Which board shapes or endpoint patterns make a two-path tile set especially brittle?
Which of these questions could become small experiments inside the editor or community grid space? A good theory board can seed concrete puzzles, counterexamples, and shareable conjectures.
localStorage.Manage grids, retrieve saved progress, and view solutions.
Choose between sharing grids and browsing the community.
Publish Repository grids publicly or manage group-only sharing from the same place.
Browse the public gallery or jump into group-shared collections from one browser landing page.
Create private groups, manage members, and publish shared grids just for those groups.
Open a group to manage members, browse private grids, or moderate it if you are an admin.
Choose between the public gallery and group-shared collections.
Browse the public gallery, test shared grids, save your own progress into Repository, and join the discussion on each card.
Open the group-sharing workspace to manage private groups and browse their shared grid collections.
All paths perfectly routed.
Configure grid → drag letters to boundaries → save to repository
Design custom routing logic for any geometry.
Drag tiles onto 🗑️ to delete them permanently.