One VCC, two jobs

The key's job

The switch joins VCC to BUTTON. While the key is down, BUTTON is VCC: your pin receives whatever voltage VCC has, with nothing on the board in between. That is the TK04's rule and it holds here: VCC must be the voltage your board's pins run at.

The LED's job

The same pin powers the WS2812. XINGLIGHT's sheet gives it a supply of 3.5 to 5.5 V, and says it is fully working from 4.5 V. So the LED would like 5 V.

On an Uno, no conflict

The Uno's pins run at 5 V. VCC on 5V gives the key a 5 V HIGH, which is what the pin expects, and the LED its full supply. Both jobs, one wire.

On a 3.3 V board, the pin wins

An ESP32, an ESP32-S3 and a Pico run their pins at 3.3 V, and none of them is built to take 5 V. Put VCC on 3V3. The key then works exactly as it should, and the cost falls on the light: 3.3 V is under the LED's 3.5 V minimum, so its sheet promises nothing.

In practice a WS2812 usually still lights at 3.3 V. Red holds up best, and green and blue fade first, because their LED chips need the most voltage. Nobody has measured this block at 3.3 V, so take that as what to expect rather than a figure. The sketches in this book start with red for that reason.

The data line is not the problem. This LED reads a HIGH on its data pin from half its own supply: 1.65 V on 3.3 V, 2.5 V on 5 V. A 3.3 V pin clears either comfortably.

Full colour on a 3.3 V board

If the colours matter, give each job its own voltage. Put VCC on the 5V pin (VBUS on a Pico) for the LED, and bring BUTTON to your pin through two resistors: 10 kΩ from BUTTON to the pin, and 20 kΩ from the pin to GND. A press then gives the pin two thirds of VCC: 3.33 V from 5.0 V, 3.07 V from an ESP32's 4.6 V. With the key up, the resistors hold the pin at 0 V.

That is arithmetic, not a bench test. The WS2812 wire needs nothing: a 3.3 V pin drives it directly at either supply.