Four steps.Every week.

Three ways to start. First: begin from a public Lichess study — paste the link, ChessNotes downloads every variation and turns them into a tree on your canvas in under ten seconds. Second: build from scratch on the board, add moves by clicking on pieces and every move becomes a node. Third: import a PGN exported from another tool.

From the canvas you can reorganize, annotate, add alternative variations. Every node has a side panel with free-form text description and, if needed, a custom FEN position. The layout updates automatically when you add a branch — no overlapping nodes, no tangled arrows.

Tip: if you don't know where to start, the QuickStart function searches Lichess studies in the most common openings (Italian, Sicilian, Caro-Kann, French…) and shows them as selectable cards. One click and they become your first repertoire.

Click "Smart Drill" in the sidebar. The FSRS-5 algorithm computes which positions are due for today's review and builds a priority queue — those most at risk of being forgotten up top. Every position is shown to you as a clean board: you make the move, ChessNotes checks and gives you immediate feedback with a green or red arrow.

If you miss, the system shows the correct move, recalculates that position's difficulty (you'll see it again sooner) and moves on. If you nail it with confidence, the position drifts into the future — days, weeks, months. Multi-move variations get drilled in full, but if you miss at move 6 the drill stops and the feedback is surgical.

The session ends when you've cleared today's positions, or when you want. Your progress (XP, streak, retention curve) updates in real-time on your Profile. Ten, fifteen minutes a day is enough to keep a repertoire of hundreds of moves alive in memory.

Connect your chess.com username on the Profile page. ChessNotes downloads your recent games — up to 6,000 by default, configurable — and processes them in the background with Stockfish 17. Every blunder (loss of 150+ centipawns relative to your ELO) is indexed with timestamp, opening, opponent and correct move.

Open the "Your Games" page and you'll find blunders sorted by severity or date. Click one, the position opens, you see the move you played and the one the engine was suggesting. You read the evaluation (cp or mate), you understand why. If it was a repertoire gap, click "Add to repertoire" and that position joins your tree.

Next time around, Smart Drill includes it in the queue. The cycle closes: real mistake in a game → enters the repertoire → enters FSRS → doesn't repeat. It's the most honest feedback loop there is: we don't tell you what to study, your own games show you.

At this point the flow runs itself. Every week: add the new variations seen in your games, drill the weak areas, correct the blunders that surface. The repertoire grows organically, calibrated on your real play — not on a GM's who doesn't look like you.

The Profile page shows the retention curve over time, the FSRS difficulty distribution of your positions, the weekly blunder count. If you see a certain opening keeps generating errors, open that branch of the canvas, study, reinforce. If a variation has had high stability for two months, FSRS shows it once a quarter — you don't waste time.

The result after three months: you enter the game with a clarity you didn't have before, at move 8 you still know what you're doing, and opening blunders have vanished. Not because you studied more, but because you studied better.