Hacker News Reader

Comment by techntoke | Hacker News Reader

This isn't self-hosted? I'm interested in a self-hosted solution that uses local markdown files and directories to manage content.

View on HN

Same, I tinkered around a bit and so far really like the semi-free form nature of how this looks and feels, plus the dirt simple, out of your way/distraction free interface is something that could easily line up with how I scribble down notes in my pocket journal, but am averse to signing up to another account somewhere for my obsessive checklisting and note-taking.

I'd gladly pay for a desktop application of this where I kept all of my notes and todos with local persistence that I could backup, migrate and move (similar to something like Notational Velocity), I like this, but I guess so far OneNote remains my personal champ in this regard-even though I have to tie it to OneDrive across my pc and surface.

Edit:

Although feverishly refreshing the thread and looking at all the praise...maybe I'll give it a more thorough shot for a couple of weeks.

This is one thing I've wanted. A local app that lets me store all my data with markdown syntax, syntax highlighting, in local files and directories, specifically for macOS although a cross platform app would be fine. One that doesn't require an account or the internet at all for it to work. I've wanted this kind of app for many years (and wouldn't be opposed to making it if no existing solution is all that great at it).

Perhaps Boostnote? https://boostnote.io/

Been using it for a work journal for a while, and it's pretty reasonable. I used to use RedNotebook before that, but I finally got fed up with the lack of Markdown syntax.

Bootsnote and Joplin works great for notes, but Joplin was less than ideal last time I checked. Hugo is really the tool I think will get me closest to what I'm looking for.

Could you expand on why it's less than ideal? I'm considering moving, so far it looks good.

Because the way it backs up your files isn't symbolic with the file system. The Hugo template I'm working on does this for you. Their are some minor annoyances, but it is the most powerful tool that I know of to get setup quickly. Wish there was some more standardization around Templates though, and the ability to do custom output templates for section lists that create paginated sections of content based on grouping, custom parameters and more. Best thing is it is all managed in a logical folder (section) and file (page) hierarchy and so far I've just about got a infinitely nested template setup with paginated sections and pages throughout, with the ability to add custom paginated sections using specific layouts for custom sections.

It's a bit less than ideal, overall. I would like to see you be able to assign a custom output format, so for each index and section kinds it would be nice to automatically create paginated sections that you want for specific layouts. I've tried this and it only sort of works, but expects them to be the pages as your lists.html or index.html, and can't paginate a path.

Also, built in client-side search. I'm working on Mermaid.js, Chart.js and Reveal.js short-code integration next.

Emacs with org mode, helm, and projectile as a starting point would hit all of your asks (although org syntax has differences from markdown, it's still quite intuitive).

I use org mode for: my tasks (gtd style), food log (using org columns), notes (easy export to any other format), habits. Alongside beorg for iOS for quick capturing on the go, I cannot imagine I would be any happier elsewhere.

Org-mode, Helm, and Projectile as a "starting point," is pretty steep. Org alone probably covers the bases.

Same here, org-mode is going nowhere, it's one of those things that just works. From wikis, todos, document authoring, agendas, code notebook, ... On mobile, there's a few options, I made one, a web app called filestash: https://demo.filestash.app/s/hn . It has a lot of the org mode candies: agenda, todos, and the real org-mode exporters: HTML, PDF, Markdown, TXT, Latex, iCal, ODT and even beamer

And literate programming too with Babel!

It sounds like a private git repo would suite you fine. You'll only need an internet connection to update when you've made changes on a different device.

I use Quiver for this purpose. http://happenapps.com/

The interface is that of a notes app rather than a board like Notion, but I just started using Joplin for this and have been really liking it so far.

I’ve been using Quiver for a few years, and love it. Totally replaced Evernote for me. Not sure if they have syntax highlighting though — there is a “code” cell, but I haven’t used it.

http://happenapps.com

I also use Quiver, I just wish I could have my notes on my Android phone. No one has yet brought together quite everything -- there's always tradeoffs between the various note apps, even though there doesn't have to be.

zim? http://zim-wiki.org/ with the "source view" plugin?

Seconded that org-mode sounds like a good place to start.

I'm pretty happy with Gollum (https://github.com/gollum/gollum/wiki) for this. Your data is just a Git repository containing a hierarchy of files. You run Gollum to fire up a local web server, which gives you a wiki interface to view and edit any Markdown files in the hierarchy. You can use a local Markdown editor if you prefer, as long as you commit your changes. You can choose whether to run Gollum all the time and expose the server to others, or just launch it locally when you want to browse your own repository. And you can use Git to create, push and pull branches.

Gollum also seems to have powerful customisation features like macros and YAML front matter, but I have yet to make use of them. For now, Gollum suits me as a simple, free alternative to Confluence. Notion is obviously a far more powerful product.

https://github.com/fabiospampinato/notable

(disclaimer: I'm building something similar)

Why is it every time something like Notion is mentioned, there is always someone who wants "local, self-hosted, Markdown"? Just build it yourself.

This clearly shows you have no understanding of the product at hand. Notion is so incredibly powerful that Markdown doesn't even scratch the surface here.

> Why is it every time something like Notion is mentioned, there is always someone who wants "local, self-hosted, Markdown"? Just build it yourself.

Why not? Is there something wrong with wanting a local, self-hosted tool? Or wanting to pay for it instead of building it yourself?

Maybe GP meant it as a call to arms? We're all here asking each other for something, but apparently none of us are building it.

Many people care about things such as "what is this company doing with the data I store with them?"

And many people don't have the time or skill to "just build it yourself," so when something awesome like this comes along that looks really appealing, you get people asking for self-hosted options.

Maybe what he meant is that he wants all of the contents and databases to reside on his machine unless an item is explicitly shared with someone else.

For some people like me that's a basic condition to even consider a tool.

Exactly... I've started building something using Hugo that does all this. That way you can create a calendar section, and then use front-matter to define the calendar entries in that section. It even creates an ICS for each item and a calendar for the entire section.

Hugo? As in the static site generator Hugo? That sounds very interesting, I've been using Hugo at work for an internal-only facing developer 'blog'.

Do you publish the progress of this project you're working on for interested persons to keep track or should we wait until you're ready to unveil it?

I've got the theme and here:

https://gitlab.com/equalos/equal

I've been trying to get it into a nice state, but it is a process.

I'd prefer a local or self-hosted product because when the innevitable "incredible journey" blog post comes out, I don't want to have to try and migrate everything to a new service. If I've paid for a product, I can at least keep using it as-is when the company goes under.

At least in Notion's case, it's a proper enterprise SaaS product with a trial tier, rather than a free consumer product attempting to go enterprise.

Proper enterprise products are hosted on-prem. This is exactly why atlassian is pretty much the only contender in a lot of cases.

I don't know if I agree with that distinction, but an on-prem version would be nice.

The problem is that economic reality doesn't allow for the market to clear for such product.

In other words: you're not willing to pay enough for such a feature to make anyone capable of building it make it for you.

There just aren't enough people like you to support self-hosted $4/month product and you're not willing to spend $2000/month for self-hosted version.

So your options are not using the product at all or using hosted, incredibly cheap, version.

Friends don't let friends use web apps. Aside from privacy concerns, relying on someone else's server functioning and being forced to use whatever the newest "improved" version are all strikes against em. Who needs another login collected by someone else (so trustworthy) that you fill with credentials that will be stolen in a hack revealed a few years down the line? There's never any real consequences for service providers. It's best not to use new web services for anything important.

Notion doesn't store password.

You can either log-in with Google account (which you most likely already have) or via 2fa mechanism where they send an expiring login link to your e-mail address.

There are no password and therefore nothing to hack.

Well... there's the data you've stored, if you're trying to keep it private. But, yes, it does invalidate the original point. :)

Having no password doesn't mean it can't be hacked. Maybe Notion won't end up in a haveibeenpwned email, but that only addresses half of a single one out of three of my objections to web applications.

You're pointing out that there seems to be this unmet demand for a self-hosted markdown based product. Rather than wonder why people keep asking for that, maybe you should include those features in whatever it is that you are building.

Maybe the demand is from the worst kind of users a business would want, those that would never convert to actual customers but want something for to tinker and install for free...

those that would never convert to actual customers but want something for to tinker and install for free...

There also might be a valid demand from those who would want to use the features of a collaboration suite like this for internal projects but find themselves beholden to internal or even client-demanded security/data retention policies.

Authentication can be implemented via a proxy, it can build very complex and encrypted data-types with pretty little automation.

Who said anything about free? I pay for all kinds of software that I install locally.

Markdown and front-matter does scratch the surface, and I do have an understanding of the product. It just isn't what I'm looking for. I care about my data and having it accessible, even outside of the application. I like organizing my content using directories for sections, and files for pages. I don't want everything saved to the cloud using a database. I want to use local files and front-matter to create content, and Hugo does this... I've started building what I'm looking for already.

Markdown is a fileformat, while Notion is an interface which can utilize this fileformat. I don't see anything in notion which I could not fit in that fileformat and some folder-structure.

The problem is that making a good interface is a though and long task. It's not done with crapping something together which then barfs out some file. That's people just can't build it themself, because it would be months and years to reach a good quality.

The demand for local&selfhosted on the other side is clear: people don't trust the companies. Companies peek into your data, sell them to other, or just disappear one day. With something under your control this will not happen.

But they're trying to run a business, so why ask the question as if it's surprising?

Why would running a business prevent them from offering a self-hosted solution? Look at Confluence, if you need examples.

Or MS Office :-B

Or every software product before about 2005.

Business is not synonymous with 'subscription model.' It's good for people to express their desire for on-prem, one-time payment products. Maybe the market can find a middle rather than the utter domination of subscription services.

It's surprising that I (for personal use) and my company (commercial use) would be willing to pay good money for the ability to run this app on-prem?

Personal data notwithstanding we have data we put in our company wiki that is simply not allowed to be hosted on cloud services without lots of red tape and compliance certifications. Offering a on-prem option, at least for us, pushes all that responsibility onto our ops team, and Notion gets paid about the same but without the hosting costs.

The problem is that what you consider "good money" is probably not even close to making such a product a viable business.

Their personal plan is $4/month.

Maybe one in a few hundred people (let's say 500 for easy math) would want self-hosted version.

Are you willing to pay $2000/month for a note-taking application? I doubt it.

Not to mention the support burden of trying to figure out why something doesn't work on your self-hosted installation.

That economic reality is why on-prem, self-hosted software is almost exclusively a very expensive, enterprise software.

Companies can justify paying a few thousand a year for installation and a salary of an in-house employee to manage and maintain it.

Individuals can't.

I'm not sure I understand your math. The support burden is very real but why would the pricing be anything other than $premium/user/month with tiers based on support requirements?

Self-hosting nerds like myself would be able to get by on a no-support 'community edition' for like $10-15/mo.

But that’s fine. My company can easily afford a few tens of thousands per month for the on-prem version.

If the on-prem version does not exist, I’ll never even try to convince them to switch, since we aren’t going to store anything sensitive on someone elses infrastructure.

So we stick with Confluence. Which everyone hates, but is still the best solution around.

what about a self-hosted gitlab instance? I can see it would fall short on a few elements but it has a great API that can be integrated with.

Check out Fossil-scm.org which supports Markdown, Fossil's variant of Markdown and plain text.

Reply on news.ycombinator.com