revise.io
The word processor engine and rendering layer are all built from scratch - the only 3rd party library I used was the excellent Y.js for the CRDT stack.
Would love some feedback!
I want to give kudos to two things:
1. It took you 10 months to build this. This is focused product development and craftsmanship which is very different from Vibe coding something. So let this be a reminder to all the "I can vibe code this or that in a weekend". Good products / experiences take time.
2. You've pursued building something in a space that anyone would normally dismiss right away: "Why would anyone use this? Google Docs/ Word etc already does this" or "MSFT / GOOG will destroy you". Good on you for picking something that is hard and building it well. I actually had this idea and almost built it but dismissed it myself for the same reasons as above. So reminder again for the builders in the back: Doesn't matter if there is a 800lb gorilla building this, if you can execute it better go for it.
Kudos!
1. How does the AI read your document? have you exposed a JSON/XML?
2. How is AI generating edits over the document?
3. How are big documents presented to AI? Was minimizing token consumption a goal?
Fantastic work, btw!
How do you know? There isn't a git repo that one can see the history of, he could have coded this in one weekend and used the rest of the time doing noncoding activities. Also, he could have made the entire thing by prompting without any hands on coding at all. The fact that it is a web app with a SaaS platform (the thing that LLM-assisted coding is the best at) doesn't inspire confidence.
That’s half of the point! Building (and selling) products requires a lot of those too.
Bugs I found:
- <tab> when in a 3rd-level indented list loses focus
- Double-click and drag gesture does not extend text selection
- Selection highlight is offset for indented paragraphs. If you select a range you can see the highlight incorrectly extended into the right-hand margin.
- Inconsistent repro: had some cases where select -> delete -> cmd-z would not fully restore my removed text (this could be my mistake)
- Toggling list style of a single indented list item can un-indent entire list, removing hierarchy; I would expect toggling to eventually return me to my original state.
- Frustration: cannot set range of indented list to ordered list without affecting all adjacent list items
- Frustration: cannot resize table rows vertically
- Frustration: on macOS, ctrl-a selects all, where the platform native behavior would be to move selection to the start of the current paragraph. ctrl-e should move selection to the end of the current paragraph, but does nothing. (macOS silently supports readline/emacs style keybinds for text editing)
I will get all of these fixed
regarding rendering/text layout, it’s all based on measureText in the Canvas API. the layout engine is my own. documents are made up of “blocks” like lists and paragraphs and each implements its own rendering, cursor movement, and layout logic.
I am going to start a dev blog where I get into the weeds. but will fix these bugs first.
if you find more like that please email me art@revise.io
Have you also considered using a solution like OnlyOffice for your product? Or a "Notion-like" lib such as Tiptap or PlateJS?
I wanted to build something canvas-based, so that eliminated most of these options. I also just wanted full control of that part of my stack... it's the core product after all. There are several TipTap/ProseMirror wrappers out there already.
You should share yours though, would be interested to see
I've been working on a ProseMirror-based editor and the biggest lesson was how much the browser gives you for free with contenteditable, even though it's painful to work with. IME input handling alone took weeks to get right, and that's with the browser doing most of the heavy lifting. Starting from Canvas seems like an order of magnitude harder.
How's the IME/CJK input working? And does it play nice with browser spellcheck?
I haven't thought much about accessibility, it's not really a priority for the project right now. If the project really gets big I might revisit that. Google docs is canvas-based, so that was enough precedent for me.
Great work pulling this off in 10 months. Good luck with the project!
It’d be very cool to have a “remove signs of AI writing” feature (based on https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing) - wishing you great success reinventing this space for the new era!
For now, the catch-all solution that anyone can attempt to use is the custom prompt, under account settings. You can instruct it never to use emdash, avoid certain cliches, or simple things like that. But you have to write it yourself for now, there's no convenient presets or anything like that.
If you have ideas please email me art@revise.io
We don’t do that with hammers, or guns. In the case of the latter, manufacturers outsource policing the thing’s use, and everyone understands that.
I also don't object to paying for something I find useful or valuable. But I really hate subscription-ware.
But I'm a local-first privacy-focused user and only by self-hosting a tool am I comfortable using it.
So for me this tool is also a no-go.
this may be something I offer one day, with some kind of one-time license. I can't really make this work without a subscription right now, because it relies on servers I host and LLMs provided by 3rd parties which have metered billing.
So, thoughts on a non-AI lightweight word processor.
If you work towards something like google docs etc., this product feels right within your category and can work with the team features at zed to a far greater degree.
Zed also natively has AI functionality so it can work for some people and the best part about Zed is that AI functionality can be toggled off too :-)
This might as well be a billion dollar unsolved problem which the team at zed could use their expertise on perhaps. Although I suggest that maybe instead of bolting these functionalities into zed itself, maybe a zed-fork can be created for a more Microsoft word alternative?
Has someone tried at making a zed extension which can somehow be a word editor or anything similar, perhaps it might be possible within the frameworks of zed now itself but I am not sure.
I hope someone at zed team reads this and solves this problem. Zed is fantastic piece of software, thanks for making it zed team :-)
It even supports code blocks, LaTeX, and Mermaid diagrams.
Also, the passive spelling/grammar checking in the editor is powered by LLMs and completely free. It will catch mistakes that other word processors won't, such as malapropisms.
Edit: Ah I see, from the OP. Unfortunately, I think Subscription-based, web-app, and vibe-coded would individually be deal breakers. Combined indicates it's not the sort of tool I seek.
Some people (myself included) will not like subscription-based, web app.
You worked 7 months on this project full time on your savings as you mention and you might've squandered any reputational gains from that with just three words and a comma.
Might as well go down in the history of hackernews but a bit negatively. I hope that you take a deeper look at how you respond online.
I have a suggestion but if you feel like you are not sure how to respond to a comment, then don't at the moment rather than typing this for example.
perhaps treat it as a learning exercise on how to answer such questions because if you ever market to anyone, customer or business. It is natural that they will ask such questions and so in a way, it might be beneficial.
Just my 2 cents.
Anecdotally, it takes a lot of patience to answer criticism in a good manner and definitely takes a lot of time to craft a good answer if you do go through that route but in the long term/even in the short term, those are some of the best messages that I have written personally which genuinely make me appreciate myself.
I wish that you can take a deeper reflection into such question as you are most likely going to be asked it quite often and having an good answer early on might be beneficial for your product.
have a nice day.
At that point, don't respond then :/
here's how I would've responded:
"Hey my project targets the niche similar to notion and others who are also web based for the most part and subscription model, reading your comments, you might be better of using non-web products and some of the suggestions in that include Qownnotes and zim. Here is a video in detail which talks about Simple, Non-Commercial, Open Source Notes for example which might help you find a solution for your needs: https://www.youtube.com/watch?v=XRpHIa-2XCE"
(Not saying that you needed to refer them to alternatives but as I knew of this video, if I were you, it might've made sense)
Another comment which doesn't talk about any other software but might've worked as well if you wished to respond:
"Hey, at some point, I do understand your take of everything turning into electron and browser-ified and software is taking 1 gigabyte of ram etc. and I understand this sentiment as well but I feel like that there is still an opportunity to make browser based subscription models towards the people who still wish for something convenient within their browser and for example, these can work absolutely great for students with their chromebooks which can have browser apps with genuine ease and targeting a large group of people to gain feedback from to hopefully improve my product even further in the future."
> Web-based and subscription payments describes the majority of software out there today
Yes and people are fed up of that too. I am sure that the alchemist isn't saying no to your product but saying yes to, for example notion. They are saying no to both and they are making their stance clear (as to why) and you are bound to have some people who are on the fence about the same thing (especially if you are launching within something like hackernews)
Edit: These are all thoughts that I can think of not even having launched anything from this product but I have sure tried my bunch of editors. I am definitely certain within your ability to write such messages as well and if you aren't, then that's completely fine too and I recommend in that case to not reply back for example.
Actually the speed is a problem when you have hundreds of pages with track changes and comments.
Maybe you should check Wordperfect or WordStar ;)
I also own mirror.forum
Good 5 letter (total word + tld) are actually pretty rare/almost exhausted now so good catch :)
The agent is model-agnostic though; I already have integrations with 3 providers. It could be extended to more free speech friendly models in the future, should there be good ones that can handle tool calls well enough.
I do a decent amount of writing on my blog and for work so I was thinking, "why doesn't this product appeal to me?"
I think I'm hesitant to spent yet another monthly subscription on something. I get decent mileage just copying and pasting sections into Claude so it's hard to justify another $8 a month on another tool.
I also do a decent amount of my editing in raw markdown files and apply styling almost as a post-process. Part of the problem is that I'm always pasting documents into corporate portals (Confluence, Wiki's, Google Docs) and they don't always copy formatting in the way I'd expect. So I just write raw text and format it after paste.
I get that subscriptions turn some people off, and I'm open to other ideas of how to make a project like this financially sustainable. I don't want to do ads :)
I think we can do much better.
The workflow of copy to chatgpt and getting feedback is just the first step, and honestly not that useful.
What I would love to see is a tool that makes my writing and thinking clearer.
Does this sentence makes sense? Does the conclusion I am reaching follows from what I am saying? Is this period useful or I am just repeating something I already said? Can I re-arrange my wording to make my point clear? Are my wording actually clear? Or am I not making sense?
Can I re-arrange my essay so that it is simpler to follow?
You can also focus your questions by selecting a segment of your document, and then writing a prompt; the agent will see what you've selected and focus its efforts on that. You can even prompt with multiple selections attached at once.
I'm hoping to add more "proactive" AI to this eventually, like automatic comments raising the critiques along the lines of these questions you enumerated. Right now the agent has to be prompted first for it to do any real thinking.
Thanks for the feedback.
You asked for feedback though.
There are chat apps that basically incorporate all the features of this editor so I am not really sure who is this for.
If this is for writing, in order to make it amazing I would personally focus on using models that are either built for that or fine-tuned specifically for writing.
Otherwise, what is the point?
Notion can do the same, perhaps more.
Notion can't quite format and export print-ready documents the same way. This is more akin to Google Docs than Notion. Revise can do proper word processor esque formatting, like page numbers, page layout, page breaks. The agent can format a paper in APA 7 with a single prompt, for example.
I already have a growing user base and customer base. People use Revise for all sorts of things, from writing stories for pleasure, to processing reports for work.
Notion is an amazing tool, but not focused on the same type of documents that Revise is focused on.
Why don't you use your local open source llm, without the interaction of big models? I mean, more work, but you don't need to pay your cut to them. Just asking.