Note that Grammarly is essentially a keylogger. The content you're typing gets sent to their servers. It doesn't work like your typical client-side spellchecker.
Their business model may be OK, it's just something that you need to be aware of (and a lot of people I've talked to aren't).
From the privacy policy:
---
We collect this information as you use the Site, Software, and/or Services:
- User Content. This consists of all text, documents, or other content or information uploaded, entered, or otherwise transmitted by you in connection with your use of the Services and/or Software. For more information about how we care for and protect your User Content, please see our User Trust Guidelines.
As an alternative there is LanguageTool[†], for which you can run your own server so the data isn't given to an extra party at all. We've been using it in [DayJob] for a while, and I use it at home too, and it does a decent job.
From my relatively experience of Grammarly (I had an account a couple of years ago) a self-hosted instance of LT is slightly better than Grammarly "free" but doesn't have the extra analysis offered by Grammarly's paid accounts.
I do this too (running it in a docker container on my server to mutualise usage across several devices) and it suffice for catching the main errors and typos. Quite happy with it.
Can indeed be recommended as substitute to grammarly free.
No doubt it works great - but I'm not convinced this addresses the claim that it's "essentially a key logger". With your suggestion, they're still getting the data.
I hope everyone assumes this by default of any web forms. I always wince a little when a colleague uses some random website to pretty-print json or encode base64. Great way to leak company internal stuff.
People who care a lot about protecting their data probably won't use it anyway. Even if your intentions are pure, your implementation could have a security vulnerability. It makes a nice bullet point feature but I'm not sure it has much promotional value beyond that.
Their business model may be OK, it's just something that you need to be aware of (and a lot of people I've talked to aren't).
From the privacy policy:
---
We collect this information as you use the Site, Software, and/or Services:
- User Content. This consists of all text, documents, or other content or information uploaded, entered, or otherwise transmitted by you in connection with your use of the Services and/or Software. For more information about how we care for and protect your User Content, please see our User Trust Guidelines.