Not an animation. A debugger you can read.
Most algorithm sites play you a movie. VizLearn lets you drive: change the input, step through it, read the exact line of code running on every frame, then flip to the written explanation when you need the why.
Code, canvas, and log in lockstep
This is the whole idea. The highlighted line, the canvas, and the log all describe the same instant: a value gets compared, the exact `if` lights up, and the log records it. When a lesson has multiple languages or approaches, the snippet follows your choice too.
Type your own input
Change the array, target, graph, or operation sequence a lesson accepts, and the animation, code highlight, and log all rebuild around it — no fixed canned demo.
Scrub, step, and replay
A real transport bar: drag the timeline, step one beat at a time, slow it down. Pause on the exact moment you don't understand and sit there as long as you need.
Hand-drawn, on purpose
Everything is sketched on a 1920×1080 whiteboard in a single hand. It reads like a brilliant TA's notes, not a sterile UI — and that warmth is what makes dense ideas approachable.
Free to start, yours to keep
Start learning with no account required. Recent lessons, bookmarks, and saved inputs stay in this browser; Lifetime syncs them across your devices.
The tools around the player matter too
The player opens the trace. The rest is for what happens next: reading the invariant behind it, or driving a structure with your own operations.
Explain is a real view
Visualize shows the trace. Explain teaches the invariant, cost, edge cases, and common mistakes without needing the player open.
Data structures take operations
Stacks, queues, linked lists, trees, heaps, and hash tables accept operation sequences, then animate the traversal and the write that changes the structure.
Predict the next step
Every sorting, search, and data-structure lesson can pause mid-trace and make you call the next move — swap or skip, go left or right. Recognizing an algorithm and predicting it are not the same skill.
The animation is the answer key
Pause on a real decision — swap or not, go left or right — and lock in your guess. Then the next beat plays, and the animation itself proves you right or wrong.
Fresh data every attempt, no answer key
Because the trace is a pure function of its input, every attempt is generated from freshly randomized data. There's no fixed answer key to memorize — each answer has to be worked out from the state in front of you.
Mastery, broken down by skill
One score hides what you missed. Four facets are tracked separately — execution trace, boundary cases, invariants, and cost — so you see exactly which part you've nailed and which still needs work.
One click to the case that matters
Most lessons carry a Scenarios row — one tap loads the input that makes a property impossible to miss: already sorted (best case), reversed (worst), a hash table that actually collides. The AI lessons do the same with their failure modes — a diverging learning rate, a retrieval query that misses, a tool call that shouldn't fire. The breaking case is the curriculum, not a footnote, and the row is free wherever it appears.
Follow feedback on GitHub
Report a problem, suggest an improvement, or request a lesson through our public GitHub Issues. Include a lesson link and screenshot when helpful, then follow replies and updates.