this post was submitted on 25 Nov 2025
88 points (100.0% liked)

Opensource

4417 readers
285 users here now

A community for discussion about open source software! Ask questions, share knowledge, share news, or post interesting stuff related to it!

CreditsIcon base by Lorc under CC BY 3.0 with modifications to add a gradient



founded 2 years ago
MODERATORS
 

What is it?

Harper is a free English grammar checker designed to be just right. You can think of it as an open-source alternative to Grammarly. I created it after years of dealing with the shortcomings of the competition.

Private

Harper is completely private, in every sense of the word.

Since Harper runs on-device, your data doesn't go anywhere you don't want it to.

That means you have 100% certainty we don't violate your copyright by training large language models.

Harper also intentionally avoids including any kind of generative AI in any part of our processing pipeline.

top 15 comments
sorted by: hot top controversial new old
[–] thatonecoder@lemmy.ca 17 points 6 days ago (1 children)

I've used it for 5 months already, at the very least. 9/10, considering how new it is.

[–] Kissaki@programming.dev 5 points 6 days ago (1 children)

Do you have a comparison to other tools like Grammarly? Were you sometimes missing suggestions or linting rules?

[–] thatonecoder@lemmy.ca 4 points 6 days ago* (last edited 1 day ago)

Haven't used Grammarly in years; however, I can tell you a couple of cons Harper has, compared to Grammarly:

1: It is English-only, for now, although it has all common English dialects

2: It sometimes gets pronunciations of abbreviations wrong, although this is getting less and less common, since the developers work a lot on it

3: There aren't plugins for it on certain editors (e.g. Notepad++), although it has for Chromium/Firefox, VS Code, among others

4: It does lack many style suggestions, but I shall reiterate that they're working on it

[–] fubarx@lemmy.world 10 points 6 days ago

Thanks! Big fan of local and on-device anything.

[–] artiman@piefed.social 8 points 6 days ago* (last edited 6 days ago) (1 children)

Very nice ill try switching from LanguageTool on EU hosted server to this

[–] Successful_Try543@feddit.org 7 points 6 days ago

FYI: You can also run LanguageTool locally or on your own server.

[–] Kissaki@programming.dev 6 points 6 days ago (1 children)

as an open-source alternative to Grammarly

intentionally avoids including any kind of generative AI in any part of our processing pipeline

Isn't that what Grammarly is all about, though? Be better than traditional spellchecking through LLM?

I assume Harper is entirely Rules based, then? Which inherently means limited to what rules where introduced manually and what the rules cover.

[–] Flagstaff@programming.dev 1 points 2 days ago* (last edited 2 days ago) (1 children)

You may like to read @thatonecoder@lemmy.ca's comment, which basically says that they're working on patching these holes.

[–] Kissaki@programming.dev 1 points 1 day ago (1 children)

That comment doesn't say anything about what I'm asking about here.

[–] Flagstaff@programming.dev 1 points 1 day ago

Fine, then: LLMs hallucinate and make mistakes. You can't beat rules in grammar; get detailed enough and everything will work. My point is that they are actively working on refining rules, which is the best any of us could ask for.

[–] BlueKey@fedia.io 6 points 6 days ago (3 children)

How does it compare to Languagetool (which you can also run locally)?

[–] moonpiedumplings@programming.dev 8 points 6 days ago (1 children)

Languagetool's browser extension is no longer open source, which has me concerned. You can still point it at a local server, but yeah.

[–] Flagstaff@programming.dev 1 points 2 days ago

Interesting, TIL, thanks. Guess I'll switch...

[–] Successful_Try543@feddit.org 6 points 6 days ago

LT supports also languages other than English, Harper doesn't.

[–] _hovi_@lemmy.world 4 points 6 days ago

From the readme:

LanguageTool is great, if you have gigabytes of RAM to spare and are willing to download the ~16GB n-gram dataset. Besides the memory requirements, I found LanguageTool too slow: it would take several seconds to lint even a moderate-size document.

Seems interesting, even if it's still early in development. I'll certainly be trying out the language server in my neovim setup anyway.