Don't know if this is in the correct forum, we are a manufacturer of engine blocks and we drill and tap a few features on this engine block. We have used thread sensors in the past but wanted to get other opinions on some thread sensors.

I’ve used eddy current in the past. When done properly we never shipped a “bad” part. Two things to watch for is what is called bad and how you zero the machine. Bad in this case is missing thread, smeared, depth of thread. It is not intended to catch where the thread is “tight” or where we’re talking pitch and functionally.
The eddy current doesn’t know a good thread. It reads a milliamp feedback and determines its variance. In essence you have to teach it what a good thread is and zero the machine. If it sees too much difference on the next reading it calls it bad / requiring a 2nd hand check. Not every part rejected is bad, just requires a follow-up check since you want false positives to some degree (error on calling a borderline part bad when it was good rather than letting a bad part being called good that was right on the line).

I had the zeroing done on every shift (for ease of consistency) but you could go a little longer. What we found though is that it is very sensitive and tends to drift over time. Thus re-zeroing to the known master. Put the master on calibration where quality checks it every so often. Yes, more than once we found the masters altered so that the machines would accept worse material. I also had red parts of a thread that were purposely made bad (minor dia too large) and saw what the milliamp difference was between the good and bad. This gave me the range to tell the machine “inside this window is good, outside is questionable”.

Hope this helps.