Trivia that adjusts its difficulty to how well you're doing.
Players and questions both carry Elo-style ratings, so each question is picked to be about as hard as you can handle. There's a daily set, a review of your mistakes and stats per category.
The app itself, running inside this page. Anything you do here is saved in your browser.
Ask it to do things
Press Ctrl+K in the app (Cmd+K on a Mac) and describe what you want. The model chooses from the app's own functions, runs them in order and tells you what changed. Some examples:
| Only history questions for now | Switches the category and picks the next question. |
| How am I doing in science? | Reads your category rating and accuracy. |
| Add a question about the first moon landing | Writes, validates and adds a question to the bank. |
Instructions you save in the same box, such as "keep it short" or "I'm a beginner", are added to every AI request the app makes.
Features
- Dual Elo
- Per-player (overall + per-category) and per-question ratings, both updated every answer.
- Target-difficulty picker that aims
- For ~60% success with a little randomness.
- Background AI question generation near
- Your rating, de-duplicated against the pool.
- 20-second timer
- Keyboard play (1-4, Enter), streaks.
- Lifelines
- 50:50 and an AI hint that nudges without revealing (assisted answers earn less rating).
- Rating-history sparkline
- And per-category ratings with accuracy.
- 36-question offline bank
- For demo mode.
- Rating uncertainty
- A deviation that shrinks with every answer drives the K-factor and is shown as a ± band, so new players calibrate quickly and settled ratings stay stable.
- Daily page
- The same ten built-in questions for everyone each day (seeded by the date, at most two per category, easiest first), one attempt, a day streak and a copyable result.
- Mistakes page
- Every question you have not yet answered correctly, with your answer and the right one, plus a retry round that does not affect your rating.
Runs in your browser
Elo math, matchmaking, timer, scoring, persistence, charts. None of this needs a network connection or a model.
Uses a language model for
Difficulty-targeted question generation with answer shuffling and validation. Non-revealing hint generation.
Notes
- Cost
- Free. Each account has a daily allowance of AI requests.
- Your data
- What you make stays in this browser on this device. Your account is only used for signing in and for the built-in AI. You can export a backup or clear everything from Settings.
- AI access
- Signed-in users get the built-in AI with nothing to set up. You can also use your own Groq or OpenRouter key under Settings.
- Source
- The code and tests are on GitHub under the MIT license. The tests also run in your browser.