Practice

25 skills. One judge. No toy answers.

1,800+ problems. Each skill is checked with its real tool: queries on real database servers, pages in a real browser, commands in a real shell, models with real Python libraries.

How it is judged

Every answer goes through the real judge.

The same path for a placement drill, a course lesson, a college test or a fest contest.

  1. 1

    You write it

    In the workspace, in your language. Run it on the sample as often as you like.

  2. 2

    We compile it

    On our judge, with the same compiler every time. A compile error is never counted as a try.

  3. 3

    Hidden tests run

    Edge cases, big inputs and time limits you cannot see, each in its own sandbox.

  4. 4

    The verdict lands

    Accepted, or exactly which kind of failure. It goes straight to your record and your report card.

While you practise

Run, get help, keep your record.

Run, then submit

Check on the sample first

Run your code on the sample, then submit for the hidden tests when you are ready.

Help when stuck

AI hints, editorials, discussions

A nudge without the answer, then see how others solved it. All switched off in contests and tests.

Your record

Insights, streaks, a profile

Personal insights on your dashboard, a streak to keep, and a public profile to show your work, or keep it private.

Start free