<p>Chaos in nonlinear systems manifests as irregular, unpredictable dynamics sensitive to initial conditions. While chaotic attractors (e.g., the Lorenz butterfly) are commonly considered in research contexts, the transient is often disregarded, although its dynamics can be complex and should be taken into consideration when the average transient time is comparable to the experimental or numerical observation timescale. Using the Hénon map as a prototype, we report fundamental concepts of transient chaos. We contrast permanent and transient chaos, quantify lifetimes through survival probability analysis, and visualize their fractal phase space distribution. The stable and unstable invariant manifolds of the map’s fixed points are introduced; these geometric structures are the backbones of the dynamics and compose a non-attracting invariant set, the chaotic saddle, that underlies the transient chaos. By analyzing how the system’s attractors change as a control parameter varies, we identify two types of scenarios where the chaotic attractor loses stability and is replaced by a chaotic saddle: via saddle-node bifurcation and boundary crisis. We also identify an interior crisis, where the chaotic attractor absorbs the chaotic saddle, resulting in an expanded attractor. We conclude by highlighting implications on other classes of dynamical systems.</p>