Duplicate in what way?

Should all the fields be checked to find an exact match? Should that include "created"/"created_at" and "updated"/"updated_at" fields? I doubt those will match exactly. How about _id, which MongoDB makes unique.

Also, if you're using something like Mongoose in Node how should foreign keys be handled when this automatic duplicate removal takes place. Should the foreign keys be corrected to the non duplicate value? If foreign keys won't be corrected automatically that would leave many records with foreign keys to non existent records, with no way to recover the record because the “duplicate MongoDB record” has been 100% removed? Populates would no longer work and your data would lose a lot of it's relationships, in the worse case scenario.

I'm not saying I'm against your idea but you need to be more clear about your explanation and expectations. What counts as a duplicate and how should related data be handled are very important questions. You can't expect very dedicated smart developers like 3T or such an intelligent community to accept such a vague and undefined suggestion.

Please give more details.