Record the outdoor temperature every minute from dawn to noon. It may rise quickly or slowly, but it never teleports from 8∘C to 15∘C without passing through every temperature in between. Graph it and you can trace the curve without lifting your pencil. That unbroken quality is continuity: tiny changes in time give tiny changes in temperature.
Precisely, f is continuous at a when three things hold: f(a) exists, the outputs approach one single value as x approaches a from both sides, and that value equals f(a). Each failure is a different break. For x1 at 0 there is no value and the outputs explode, an asymptote. For x−1x2−1 at 1 the outputs approach 2 but there is no value, a hole. For a taxi meter that adds $0.50 every 200 m, the two sides of each jump approach different fares.
Being defined everywhere is not the same as being continuous: the taxi fare has a value for every distance, yet its graph is a staircase of jumps. Smoothness is a separate matter too. ∣x∣ has a sharp corner at 0 but no break, so it is continuous there, though it is not differentiable.
Think of it like this. A continuous function is like a hiking trail with no gaps: to climb from 200 m to 500 m of altitude, you pass every altitude in between. A trail with a missing bridge has a break where you would have to leap. A trail can still turn sharply and be unbroken, just as a continuous graph can have corners.
Core idea. A continuous function has no jumps, holes, or breaks; its limit equals its value at every point.