This is a 2D companion to the site's 3D Van der Pol oscillator. Instead of one trajectory in a rotatable scene, it integrates the exact same second-order ODE — ẍ − μ(1−x²)ẋ + x = 0, rewritten as ẋ = y, ẏ = μ(1−x²)y − x — with real fixed-step 4th-order Runge–Kutta, run simultaneously from five different starting points scattered around the (x, ẋ) plane. Watching all five spiral onto one shared closed curve is the most direct way to see what a limit cycle actually is: a single attracting trajectory that "forgets" where it started. A separate x(t) panel tracks the first trajectory's real waveform over time, so the shape change driven by μ — sinusoidal at low μ, a sharp relaxation oscillation at high μ — is visible as an actual signal, not just as a loop shape.