The Grammar feature seems to have weird suggestions/cycles too on a little bit more testing. Curious to see how this improves. A local only, one-time-purchase grammarly alternative is appealing!
[1] Loosely; I’d say not referencing any networking entrypoints or dlsym() also counts, as working around that would be very non-deniably malicious.
Edit: Alas, Hacker News also removes the extra space after periods.
Two spaces after periods is a kludge invented for typewriters that had monospaced fonts and touch typing teachers need to stop teaching it in the modern era where most writing uses proportional fonts.
You can try it out here:
https://chromewebstore.google.com/detail/grammit-the-ai-gram...
LLMs can also correct other sorts of mistakes, such as correcting "The first US President was _Ben Franklin_" to "George Washington".
And they can also cause new factual mistakes that a grammar checker never would!
Just have a Windows Scheduled Task kick off this bat file:
SET PATH=SET PATH=C:\program files\Amazon Corretto\jdk17.0.15_6\bin\
start javaw -cp languagetool-server.jar org.languagetool.server.HTTPServer --port 8081 --allow-origin "*" -l en-US --languageModel "C:\LanguageTool\LanguageTool-6.3\ngram"
Does not catch a singular/plural discrepancy between the subject and the verb in a sentence--a common mistake when the expressed thought applies equally to one thing or to many things.
Does not catch a missed indefinite article--a common mistake for speakers of languages that don't have articles. Similarly, does not catch the use of the indefinite article for a thing already mentioned before.
Does not even catch the obvious "don;t" typo.
I use Harper in my Neovim setup and its been great. I just add things that are missing from the dictionary when I come across them in my codebase
The fluency suggestions are seemingly largely malfunctioning. It frequently suggests starting and ending sentences with quotes, although it also makes some useful suggestions. There seems to be an issue with analysis running synchronized, or something like that; when I type into a text field and Refine starts to run, it often blocks text entry. Selecting a suggested replacement blocks the app for half a second or so. Neither of these problems occurs with Grammarly or Language Tool. I also noticed a bunch of issues that Grammarly catches (like verb agreement) that Refine does not.
But this is an amazing first release and extremely promising. Congrats!
I realize I'm a niche :)
Does anybody know of such a tool?
Two of your 4 questions were answered in the first content block
This phrase is offensive and violates my safety guidelines. Therefore, I will not revise it. I am programmed to avoid generating responses that are obscene, or that contain profanity.
unsloth/gemma-3n-E4B-it-GGUF
You are a precision editor guided by a custom style manual. Your tasks are ordered by priority.
Your primary rule is to consult the provided REFERENCE DICTIONARY. Any term on this list is correct and must be preserved exactly as written.
Your secondary rule is to refine phrasing and sentence structure to improve clarity, conciseness, and flow. The goal is to make the text read more naturally and professionally, while **strictly preserving the author's core meaning and tone.**
Your final rule is to output ONLY the clean, revised text. You MUST NOT add any commentary, greetings, or explanations.
REFERENCE DICTIONARY:
{{dictionary_words}}
Revise the following:
"{{sentence}}"
You are an expert editor. Your single most important goal is to improve the fluency and clarity of the following text while STRICTLY PRESERVING the author's original voice and meaning.
You MUST follow these rules:
1. Only rephrase sentences that are genuinely awkward or unclear.
2. Never make changes for purely stylistic preference.
Return ONLY the clean, revised text.
Revise the following:
"{{sentence}}"
{{dictionary_words}}
You are a silent grammar correction engine with a custom style guide.
Your primary rule is to consult the provided REFERENCE DICTIONARY. Any term on this list is correct and must be preserved exactly as written.
Your secondary rule is to correct all other grammar, spelling, and punctuation errors in the main text.
Your final rule is to output ONLY the clean, corrected text. You MUST NOT add any commentary, greetings, or explanations.
REFERENCE DICTIONARY:
{{dictionary_words}}
Correct the following:
"{{sentence}}"
You are a silent grammar correction engine. Your sole function is to receive text and output the corrected version. You MUST NOT add any commentary, greetings, or explanations. You will only return the clean, corrected text.
Correct the following:
"{{sentence}}"> Sometimes I still make mistakes with articles and prepositions, but my grammar is getting better every day I practice.
In American/Simplified English, this is grammatically correct. However, in 'full fat' English, practice is a noun, whereas practise is a verb; e.g.:
> I go to my practice to practise medicine.
The problem I have with this website is that it's entirely concerned with peripheral issues. The product respects my privacy - good. The product is performant - good. The product doesn't require an Internet connection - good. The product works in many writing apps - good. The product has transparent pricing - good. But I don't give a shit about any of this until you convince me that this will consistently do the correct thing, and this website singularly fails to achieve this.
> ... in 'full fat' English ...
English is a bastard of a language and getting messier every day as new nations adopt it is their standard language.
Setting the bar where it is well written and unambiguously understandable is IMHO completely fine for a 15$ product.
Having a text spell checked to comply with contemporary Oxford English is likely not the goal of this product.
I disagree strenuously with this idea, because it suggests that there is one 'big' English in which anything goes. A better idea is the one of the register[0]: there are many Englishes, many sets of rules. Different rules are used in different regions, by different groups of people, and have different connotations (e.g., the King James Bible was intentionally written in a form of English that was considered archaic at the time because that would make it sound more grandiose).
If I were to use this tool, I'd be using it to ensure that whatever I'm writing is well-received by my intended audience. Because English usage is so varied, I would need to be able to control the register that it uses to ensure that the output is suitable. The fact that the product website doesn't even mention a list of supported languages, let alone supported dialects and registers within those languages, has a very everyone can see what a horse is kind of feeling[1].
[0] https://en.wikipedia.org/wiki/Register_(sociolinguistics)
Absolutely, but try to make a run-of-the-mill LLM understand this.
If you chose Oxford because of the Oxford English Dictionary, note that it's not regular en-gb, it's en-gb-oxendict. "the OED often favo[u]rs "-ize" (and its derivatives) over "-ise" for words derived from Greek roots, and may also include historical or less common usages."
What you suggest does seem like a good early doors feature; but the cut they've made seems to be the right one to prove market potential.
Perhaps you live under a rock, I don't really know, but it maybe just happens that you're not a target audience? There's for sure many folks, and even more so companies, who will value their own privacy more over practice vs practise bikeshedding.
British English is still influential over most of the Commonwealth, ie a large number of countries.
If you use these words in writing for Americans and you are not a citizen of the British Commonwealth, you instantly mark yourself as arrogant.
It never occurred to me that this could ever be perceived as arrogant (even if only when referring to someone with a different background.) And I wouldn't have thought it would mean anything more than a certain language cosmopolitanism, lah ;) (Hope that joke comes through! It's been decades since I had much exposure to Malaysian English.) Can you explain why this might be, please?
British Commonwealth authors (well, really any author I know to be not from the USA) get a pass because these are the spellings they were taught. Nothing wrong with that.
This is a phenomenon I've only noticed in the last two decades or so. I don't know if American students are now being (wrongly) taught British spellings in school or they merely think their writing will carry more weight if it has a British "accent" but it just seems arrogant to me.
The OED is a useful resource but it is not our dictionary of final arbitration. Americans should use the American Heritage Dictionary.
As an American English speaker, I have in the past used UK spellings when communicating at work with a group that I know only contains British English recipients. There is nothing wrong with that -- anything that makes communication more fluid should be welcome.
I believe the arrogance angle exists in a situation where an American English speaker with no British English education is using British spellings when communicating with other _American English_ speakers to purposefully create an air of superiority. If you do this, even if no one says anything, they definitely notice.
For other English dialects, my personal take is that most Americans (at least the ones who travel or interact with foreigners personally or at work) will assume they either are or are heavily influenced by British English due to history.
Non-native English speakers don't learn English exactly the same way as natives learning from their parents.
What do you regard as the British Commonwealth, and what variant of English spelling to you expect people to use who are not part of this grouping?
Just so that you know, in case you don't, "US English" is used just about exclusively in the US, and UK English is used in most of the rest of the world, despite the fact that most of our devices incorrectly default to US English.
To be sure, I don‘t have any dog in this fight, just highlighting a fact from my experience that many here might not know.
Nobody actually uses UK English here, except for English teachers. Computers don't. TV doesn't. Corporate jobs don't. And so regular people don't either.
Without such a conscious choice, yes, Americanisms do seem a fair bit more globally pervasive and easy to fall into "by accident".
> Without such a conscious choice, yes, Americanisms do seem a fair bit more globally pervasive and easy to fall into "by accident".
You'd need choice and enforcement - unless such organization is testing for Received Pronunciation during interviews and filtering out people who cannot into Queen's English[0], I'd wager most of the members in such org, who don't come from UK (or a few related countries), will be speaking "British English" with distinctly US pronunciation. Because while an organization can make a conscious choice here, for most people, learning a second language is a long-term endeavor that largely happens "in the background", and it's very easy to learn a blend, with UK English being present in schools, and US English everywhere else.
--
[0] - Or is it King's English now?
Non-native US-English speakers are not viewed in the same light (in my opinion).
That is simply not true. Maybe that's how you feel about it, but it is generally not the case (or even considered, really).
The reference to Brian being in the kitchen is understandable only for Gen X in France. (https://fr.wikipedia.org/wiki/Where_is_Brian%3F, only in French unfortunately but you also have https://youtu.be/11jG7lkwDwU?si=gQoVL_AmBfQMrZbX&t=57)
Correct: practice is a noun, and practise is a verb, in non-US English. I don't have my (twenty-volume) copy of the Oxford English Dictionary to hand, but Wiktionary has an explanation under 'usage notes': https://en.wiktionary.org/wiki/practise
I see.
In the UK, it's considered good form to be humorous when making an argument; I gather that in the US, you're supposed to sound like you're making a speech to prepare troops for war. I apologise if, in the course of describing how a product is unsuitable for use in my culture, I made that argument according to the norms of my culture.
Banter (Simplified)
Ah well, I once read an argument for “EU English”. If it’s anything like my Dunglish (Dutch-English “What talk you about”) it would indeed be simplified.
> The Simplified Spelling Board was announced on March 11, 1906, with Andrew Carnegie funding the organization, to be headquartered in New York City.
Some big names here.
> The New York Times noted that Carnegie was convinced that "English might be made the world language of the future"
He wasn't wrong.
> and an influence leading to universal peace,
That's still to be seen.
> but that this role was obstructed by its "contradictory and difficult spelling".
Well now.
It's interesting to scroll through the list of proposed changes; 100+ years later, many of them seem to be the default/correct spelling, but just as many look wrong, even when following the same transformation rules. E.g.: "brasen" -> "brazen" vs. "surprise" -> "surprize".
Thanks for linking this!
I don't know if the parent comment was trying to equate American English and Simple English - I can see it as a way to dismiss American English as a "lesser" language (which it isn't, as you say), but I wouldn't start by assuming that.
Weirdly (to my brain), Americans always spell practice with a C, but always spell license with an S.
> OpenHunt tried solving a problem for the content makers without providing any additional benefit to the content consumers. It's a nice, heart-warming mission. But in the end of the day, content is king, that's what consumers want.
> There have been many examples of people rallying around a "free and open" version of a service. They fail to realize that the end consumer barely cares. Look at voat (Reddit), app.net (Twitter), Diaspora (Facebook), even ycreject.com (Y Combinator) tried to be a thing for a while.
> If someone is able to make it "free and open" while also making it a better experience than the alternative, then it'll be a big success. But so far everyone gets that wrong.
The trick is getting the content creators there, but most of them are ultimately and fairly interested in making money, and your new platform wont have that for them.
Bluesky has done alright, but that was a black swan event Elon Musk inspired.
Couldn't American schools teach English the same way, like, that there are no single proper way ?
I worry that this will make my writing more likely to fail an AI coursework detector, which could really impact my life. The risk just isn't worth it till someone has tested the output through all the big players (turnitin etc.)
Colour and licence are so quaint.
I could cite many sources but https://en.wikipedia.org/wiki/American_and_British_English_s... provides a simple view.
But when you look at things like the type of English being taught in school, or the language used by the government, it'll be UK/International English.
Commonwealth countries are relatively minor in population compared to the USA and the rest of the world.
And yes, default US spelling in devices does play a role. See prev. point about technological and cultural dominance.
Given India's presence in the Commonwealth, this seems an odd assertion. 2.6 billion people live in the Commonwealth[1].
[1] https://en.wikipedia.org/wiki/Member_states_of_the_Commonwea...
We have enough American cultural hegemony as it is. It frustrates me no end that I regularly am unable[1] to configure software to use my preferred version of English.
[1] or it's extremely difficult. CLion I'm looking at you; every available option I have is set to British English, but still you insist on telling me my Colours should be corrected to Color etc. :(
You don't need docker (a Linux-only piece of tech) to run a java application. Even though I'm on Linux, and docker experience is waaaaay better here than on any other platform, I wouldn't in my life consider using it to run LanguageTool.
java -cp languagetool-server.jar org.languagetool.server.HTTPServerTheir recommended process is :
1. Install fasttext (it's an official Debian package, but you have to compile it on Windows).
2. Download and uncompress the LanguageTool release.
3. Create a config file.
4. Launch the server with the java command (of course, a JRE must be installed).
5. Connect to the API, e.g. with the browser extension.
Running a ready-made docker image replaces steps 1-4 and removes the need to install Java globally. Some will prefer it this way.
> Running a ready-made docker image replaces steps 1-4
We can go pretty low level in the docker option too.
1. Download a Linux installation image
2. Download a hypervisor
3. Install Linux on the virtual machine
4. Install docker in a virtual Linux machine
5. Launch LanguageTool container
6. Configure networking between the host and the container in the guest
7. Connect to the API, e.g. with the browser extension
Obviously I initially oversimplified by omitting the configuration step. But adding download step to inflate the complexity is not a fair play ;-)
I'd argue that running a platforn-native artifact is both simpler and easier than involving virtualization. Even if steps 1-4 are done by some magical tool like rancher or docker-desktop
You don't need to fight with installing (the correct version of) java. I wouldn't in my life consider installing java on my OS directly.
For anyone with docker already installed (most people these days)...
git clone git@github.com:gardner/LocalLanguageTool.git
cd LocalLanguageTool
docker compose uphttps://github.com/sonnyp/Eloquent
It is restricted by Flatpak (i.e. Control Groups and Namespaces):
* No file-system access
* No access to devices
* Network Access is allowed (API of Languagetool is only reachable via REST?)
PS: LibreOffice had to fight years to remove Java which plagued the project.
Does Refine solve this?
> Can’t find the answer you’re looking for? Feel free to sent us an email at support@refine.sh
Should be "send". Especially with this kind of product, you shouldn't have grammar mistakes on the website :)
Gemini subscription costs as much, allows me to do much more, and I can call it from vim or emacs without some arcane acrobatics
They don't advertise this because they are trying to push their paid online services. I have complained about this, but they didn't seem to care.
Legally, there is no entity behind that responsible for privacy (1), and honestly, I don't see even minimal reason to trust this software from a legal perspective.
codesign --display --verbose=4 Refine.app 2>&1 | grep sandbox
Personally, I only use software that was either built on my machine or downloaded off of the Mac App Store (MAS apps have the be mandatory sandboxed).
[1]: https://developer.apple.com/documentation/bundleresources/en...
Many popular apps, like Grammarly and Raycast, are also not sandboxed for the same reason. Without these APIs, Refine wouldn’t be able to provide a seamless text refinement experience. Otherwise, why wouldn’t we just use ChatGPT to correct the grammar?
Is your objection that he is distributing this by himself, instead of through the app store? Or that it appears that he is doing it as an individual instead of a company?
Sure, it's a little sketchy, it's a guy with a website and a privacy protected domain and that's about it. But if anything were to happen you would be suing the developers of refine.sh.
I guess I do see your point though. For my software I have indeed created a legal entity and can be easily looked up.
[1]: https://truelist.co/blog/linux-statistics/ https://gs.statcounter.com/os-market-share/
Love the pie graph that that adds up to 153%, though.
gbalduzzi•6mo ago
runjuu•6mo ago
In upcoming releases, Refine will support custom prompts and BYOK (Bring Your Own Key), allowing you to use any large language model you want.
pyman•6mo ago
You do realise you're already using an optimised model built for everyday devices, and that model includes some serious innovations in parameter-efficient processing, right?
You're a great developer, and it looks like you're thinking about adding features like BYOK quickly to please more users. But in doing that, you might be missing the real innovation you've already created. You've basically built a version of Grammarly without the privacy issues that make most IT departments ban it.
No one wants Grammarly or your tool sending corporate emails or documents to a language model. Privacy is what sets you apart from Grammarly. It's your biggest feature right now. Add a big table to your site comparing your products privacy with Grammarly's. That's your strongest selling point, and probably the only feature that can truly compete with the big players.
My advice? Keep improving the app, but keep the model local. Keep it private. That's the killer feature you've got.
rubslopes•6mo ago