Slitherlink
Also known as loop the loop, fences, dotty dilemma or the number loop puzzle.
Connect the dots into one continuous loop that never crosses or branches. A number tells you exactly how many of that cell’s four sides the loop uses — a 0 means the loop stays away entirely, a 3 means it hugs three sides. Every puzzle here has been counted by a solver that shares no code with the puzzle builder, and only puzzles with exactly one solution are used. You never have to guess and hope.
Daily Puzzle
Settings
Tap an edge to draw a line, again for a cross, again to clear. Drag to keep going.
How to play Slitherlink — the rules
- Goal: draw one single closed loop along the grid lines. It never crosses itself and never branches.
- Tap or drag along an edge to draw a line. Tap again for a cross — your own note that no line goes there — and once more to clear it.
- A number says exactly how many of that cell’s four sides the loop uses. Cells without a number can have any count.
- Every dot has either no lines or exactly two. Three lines meeting at a dot is always wrong, so we mark it in red.
- One loop, not two. Finishing with two separate rings satisfies the numbers but is not a solution — the check requires a single connected loop.
Keyboard: U or Backspace to undo.
What the badge means: before a puzzle ships, a solver that shares no code with the builder is given nothing but the numbers, and counts how many single closed loops satisfy them. Only puzzles that come out to exactly one are used. If the count cannot be finished within the search budget, the puzzle is discarded rather than assumed unique — an unfinished search proves nothing, and we will not call it a guarantee.
Strategy and tips
Frequently asked questions
What is Slitherlink?
A grid of dots with numbers in some cells. You draw a single closed loop along the grid lines so that each number matches the count of loop segments around that cell. It is also known as loop the loop or fences.
Is there always exactly one solution?
Yes. Before a puzzle is used, a solver that shares no code with the builder receives only the numbers and counts the solutions, stopping at two. Only puzzles that come out to exactly one are kept.
What if the solver cannot finish?
Then the puzzle is thrown away. A search that ran out of budget has not proved anything — it might have one solution or a hundred. Treating "we do not know" as "unique" would make the badge a lie, so we discard instead.
Do I ever have to guess?
Because the solution is unique, every step you can justify is safe, and there is always something to deduce. What we do not promise is that every deduction is easy — the harder sizes may need you to look several moves ahead.
Why are the puzzles made in advance?
Because proving uniqueness takes a full count, and removing each number means counting again. A ten-by-ten board takes tens of seconds to build — far more than is reasonable to make you wait for. So we build them ahead and ship a collection that has already passed.