This question already has an answer here:
(Modified as per request)
What is the upper bound for whole-body adaptive regeneration?
This is not about how fast a cell can regenerate, but about whether the cells could house sufficient technology to effectively rebuild part or all of a person, regardless of timeframe.
Individual cells matter only in that you have to know enough about each cell to know when it is damaged, how it is damaged and whether you want to keep it. You don't want a cancerous tumour regenerating when you try and kill it, but you do want accidentally killed cells repaired and you do want the original cells in the cancer replaced.
Cells aren't enough. A knife to the heart, if it's sharp enough, could theoretically separate cells without killing them. The regenerative system has to recognize this and repair even though it's not cellular damage.
If cells go from oxygen starvation to oxygen overload, the mitochondria will shut the cells down. Again, not physical damage but needs fixing.
Nor is this about trivial replacement, but adaptive regeneration. If you're killed by some deadly toxin, there's not much point healing only to die again. So you modify the DNA, as noted below, to make the same form less likely.
Cells have numerous tiny cavities, vacuelles, which may once have housed DNA. Cells also generate an electrical field, which you could potentially modulate.
My main interest is in whether this is sufficient to house cooperative nanobots and a distributed computer for whole-body adaptive regeneration (ie: you change so as to oppose the cause of death, so don't get killed again), but there may be all kinds of other uses such as the body becoming a computer interface to the nervous system.
Factoring in that DNA is self-modifying, which complicates things, and the inherent simplicity of nanobots, there are limits.
On the other hand, with suitable error correction codes and suitable algorithms, if you could dismantle a cell and then rebuild it, shift calcium molecules, alter DNA, etc, according to the results of the computer, then the processes aren't impossible. Size constraints and energy budgets might make it impossible for humans, at least beyond a certain level.
So, the question is, what is that level?