TP4056 units have a 1% variance. Which means it has a range of +/- 0.042V, or 4.242 - 4.158V. Pretty large range, imho. However, you can get a batch of the boards, then test each one with the same cell (this is pretty critical to the calcs). You only need to slightly discharge the cell between tests, say to around 4.0V (the TP unit may not trigger to charge at 4.10V).
The other option is to go with a slightly more expensive, but far more accurate charger, the TP5100.
Not only is this a more accurate charger, it has a wider range of input (4V - 15VDC), and there are far fewer fakes out there than the TP4056 (fake ones have >1% variance, can get extremely hot during charging, charge >500mA regardless of the resistor used).