Getting started

Katabench works in a browser. You do not need to install an SDK, clone a repository, or configure a local database to complete a puzzle.

You can explore as a guest, but create a free account if you want solved progress, submission history, learning-path completion, XP, streaks, achievements, and rewards to follow you across devices.

Pick a puzzle

Open the app and choose a puzzle from the catalog. Puzzles are grouped into seven tracks (Algorithms, Database / EF, Refactoring, Architecture, Secure Coding, Design Principles, and Test Writing) and rated Easy to Hard. Every puzzle states the goal, the constraints, visible examples, and the bar you are expected to clear.

For the simplest first experience, choose Tour in the puzzle list and start its first Easy exercise. Choose All puzzles when you would rather browse by track. You can search by title, summary, or category and combine that search with difficulty and solved-status filters.

Locked rows belong to a plan you do not currently have. A Free badge inside a locked track marks a preview you can solve without upgrading.

Read before you type

The left problem panel contains the required class and method, constraints, examples, sample-test count, and hidden-test count. Keep the starter signature unchanged so the grader can call it.

Progressive hints appear below the brief when the puzzle includes them. Database exercises can also show their schema, relationships, and sample rows. Use these as part of the problem statement, not as an afterthought after a failed submission.

Run, then submit

There are two very different buttons, and the distinction matters:

• Run executes only the visible sample cases (plus an optional custom input you type yourself) and shows you the return value and your Console.WriteLine output immediately. Run does not count against your submission quota. Free has a separate daily run allowance; paid plans lift it under the current plan terms.
• Submit grades you against the full hidden suite, including the large performance tests. Submissions are what count toward your daily quota on the Free plan.

A good rhythm is: Run until the samples pass, try at least one awkward custom input, reason about edge cases and complexity, and then Submit.

Use Ctrl/Cmd + Enter for Run and Ctrl/Cmd + Shift + Enter for Submit.

Reading your results

Use the Output tab for visible cases, return values, console output, custom input, and expected-versus-actual differences. Use Tests for the quality scorecard and the complete graded evidence.

Each applicable test reports pass/fail, server-measured elapsed time against its time budget, and allocated bytes. Failures on visible cases show a character-level difference between expected and actual output. Hidden tests mask their input and output; you see their name, status, timing, and applicable quality signals.

Different tracks add different evidence. You may see query-plan rules, code-quality metrics, design rules, adversarial security checks, or planted bugs caught by your test suite. Read How grading works when a result dimension is unfamiliar.

Stuck?

Start with the progressive hints. If you are still blocked, every puzzle has a Reveal solution option: load a correct reference into the editor, study it, then Reset and rebuild it from memory. Revealing does not mark the puzzle solved and does not consume a submission.

Your progress

Solved puzzles are marked with a check. A clean Submit also updates learning-path progress and, for a signed-in learner, activity and rewards. Use Next puzzle to continue the current path or move to another useful unsolved exercise.

Your in-progress editor buffer and custom input are preserved per puzzle in the current browser. Signed-in solved progress and activity follow your account. See Finding your way around for the rest of the app and Using the puzzle playground for every workspace control.