Ultimately, the issue presented is as much a technical one as a legal one.
You can protect software written in a programming language that you don't own the rights to via copyright. Indeed, the vast majority of copyrighted programs are written in programming languages not owned by the author of the copyrighted work.
You can't meaningfully protect software that is subject to a creative commons or MIT license for commercial purposes.
The question then becomes, is the material you want to protect simply software written in an open source programming language, in which can you can protect the software, but not the underlying language, or is it a mere implementation of open source software that is not so transformative that its character as open source software (or a derivative work of open source software) is overcome.
This would be a question of fact for a fact finding at a trial and would require considerable technical expertise and understanding to evaluate.
I know so many startups and platforms are written with open source
software so I'm sure there is a clear answer to this but I can't seem
to find one on the web.
There aren't a lot of clear precedents governing where the line should be drawn. Software of any kind you would recognize as such has only existed for about 50 years. Open source software has only existed for 20-30 years depending upon how you count it. And, open source software has only had widespread commercial use for an even shorter time period.
This is an incredibly short among of time in terms of legal history, and it doesn't help that business to business copyright litigation doesn't take place at all in state courts and makes up a pretty small share of the overall federal court docket. And most of the copyright cases that are brought in the federal courts are very simple ones. For example, as of the year 2015, most copyright lawsuits in the U.S. merely alleged that anonymous Internet users downloaded pornography without permission to do so:
[T]he adult website Malibu Media is a prodigious enforcer of its
copyrights. According to law professor Matthew Sag of Loyola
University in Chicago, Malibu alone was responsible for nearly 40
percent of all copyright filings in federal court in 2015. Litigation
against anonymous downloaders, by Malibu and other copyright
enforcers, made up nearly 60 percent of the federal copyright docket
last year.
In 2016, there were fewer than 4000 copyright infringement cases filed in the entire U.S. (in all media). And, the percentage of civil cases that go to trial at all in the federal courts (rather than being resolved by a settlement or default judgment) is very small to start with, with many of the cases that are resolved on the merits not appealed. There are fewer than 400 appeals per year, nationwide, in copyright, patent and trademark cases combined, and patent and trademark cases make up and outsized share of that total because the amount of money at stake is higher on average in patent and trademark cases than in copyright cases, and many copyright cases involve copyrights in media other than software.
Maybe there are a dozen or two software copyright appeals a year these days. The figures from the last few years, however, are much, much higher than they have been historically, particularly in the area of software copyrights, which is why the case law is so thin on so many more sophisticated areas of software copyright law.
GPLv2, which is a lot more restrictive and may require you to publish source code.