> A simple blogging platform I built out of frustration with other tools
I like this phrasing. Many of the best things are made out of frustration - it might be one of the most important building materials for software.
‘Cool blogging platform! What’s it built out of?’
‘Frustration’
Definitely better than building one out of fear, anger or greed.
Actually, in some seriousness, I think stackoverflow was built out of anger - and used that to clearly define what their differentiators and value proposition were. So maybe hanging on to that original motivation could be an important part of setting out what makes this different.
So true. I built https://www.thingybase.com because I got pissed at how stupid hand held label makers are. “I should be able to use my voice or a real keyboard from my phone to label stuff”. A few years later and I’ve got a full blown home inventory system.
Appropriately, ODD is an acronym for Oppositional Defiant Disorder, “A disorder in a child marked by defiant and disobedient behavior to authority figures.”
Hi! And thanks :) Not sure if frustration is the right word (not native speaker) but I wasn't happy with existing tools for a reason or another so I thought why not build one that I truly like :)
Hi, congrats on launching! I can see that this project will be giving quite a few companies a run for their money!
I was curious about your email-as-a-CMS solution. How does it work? Did you take inspiration from an existing project that does this?
I am aware of how Amazon uses something similar to allow people to email PDFs to their Kindle devices. But I am very interested in how you got your solution to work.
To be honest I borrowed the idea of publishing via email from Hey World. That platform is too minimal but it's nice that you can post also by sending an email, so I implemented this feature in DynaBlogger right away because I think people might like it :)
The way it works is very simple. Each blog gets a unique email address, you or another user with access to the blog sends an email to that address and the post is published immediately. Images and code blocks are supported. The format is expected to be Markdown, so it's best to send the email as plain text to avoid complications depending on the email client.
For now the editor in the app is Rich Text only but I am planning to add Markdown there too if I see it requested. :)
Did you implement the email feature in DynaBlogger from scratch?
I have been on the look-out for open-source projects that do email-as-a-CMS for a long time. I build websites and blogs for companies and individuals, and any CMS that I include (other than self-hosted Ghost) is very harsh on the content teams of these companies.
They are non-technical people who don't understand frontmatter and markdown. Email looks like the perfect publishing medium for this demographic.
Can you give me some direction on how I can build myself an email-CMS solution similar to what you have implemented in DybaBlogger? Thanks!
Sure! DynaBlogger is a Rails app, so I could use some built in support for this called ActionMailbox. It's pretty simple really and it supports various email services. I use Postmark for this :)
The email option is a simple concept, and it's the simple stuff that we often look over. I'm a big fan, nice job!
$8 a month seems like a great price point for a fully managed service. Just out of curiosity, what's the plan if someone's blog gets a lot of attention and ends up using quite a bit if bandwidth? I don't think that'd be an issue unless there are a ton of images, but I'm always interested in how devs choose to handle these edge cases.
Hi! Thanks for the nice words! :)
For now I am not limiting traffic etc because I'm still in the beginning so I'll see how it goes. I have a ton of bandwidth paid for with the servers so it will be mostly a matter of load on the servers in the end. But for now the Kubernetes cluster costs me just 70e/mo with quite a nice amount of reasources. It's cheap because I manage everything myself.
I can scale with one click, but if I find in the future that some user use a lot of resources then I will implement some limits and perhaps more plans. For now I want to keep it as simple as possible :)
My blogging engine [1] also supports submission via email (and has for twenty years). I have a unique email address for it and it has never been spammed (or received anything from anybody except for me). The only check my software does is check the author field (in the body of the email [2]) is allowed.
Evernote has a really cool email-to-note feature that relies on a secret custom email address. The additional requirement is that only the email addresses connected to the Evernote account can be used to send information into the system.
I'm not sure how they do this under the hood, but if anyone knows, please let us know.
The Evernote feature has an additional layer of awesomeness... As it is mainly used to email items into a specific Evernote notebook, you can choose which notebook to send this magic email to, by adding @notebookname to the email subject. Hashtags that follow the notebook are added to the note as well.
Dynablogger is really cool! I'm signing up now and look forward to diving in. Great work!
At the moment authorization relies on the secrecy of the unique address, yes. I wanted to keep it simple. As long as you don't share that address, which is close to impossible to guess, you should be fine. Is there something else you would recommend I implement, that doesn't impact on usability? Thanks!
I'm not an expert here, but I suppose that only processing emails that have pass all the SPF/DKIM/DMARC checks would help. This may be something your email server already does, before your app ever sees the emails.
Otherwise, requiring a digital signature doesn't seem like a huge usability hurdle to me, most email clients support this pretty routinely.
Yeah I think it would be a little overkill. The unique email addresses are almost impossible to guess so you just need to make sure you don't share them with anyone.
You would provide this file for them to download alongside where you show the email address; the file would be equivalent to an API key - should be a symmetric key. It's sort of just swapping out an Authorization header for a file with an expected checksum/contents.
If the barrier is technical abilities, trying to set up public/private keys and/or signing the emails could hinder adoption.
One more suggestion that would, IMO, be fairly simple. Tell the user to write in their subject line a {accessKey: 'someKey'}. The access key is given to them at the time of creation and can be regenerated as often, or rarely, as the user prefers. If they want update it weekly to provide an extra layer of security they can, or they could keep it the same forever with it just bein a handy backup in case they do link the random email. Just a thought.
How about having an option to “approve” the post. For example, after I email the post the system sends an email to my registered address with “publish” and “deny” buttons. One click and it’s done.
I'm taking a note. Shouldn't take too long to implement but I think this would be optional as some people could be happy with the just difficult to guess address. :)
Maker here. I've made a ton of improvements (like, a lot) over the past 8 months and I'm excited to share an update with you all on DynaBlogger.
As a web developer, I very often find solutions to problems in blogs, so I like to give back by publishing a post whenever I find a solution to a problem on my own or learn something new.
For my blog I have used many tools over the years, from a heavyweight CMS like WordPress to static site generators and everything in between. I do prefer using a CMS for this, so I can just edit a post, hit publish and be done with it, rather than writing a post locally and pushing to a repo waiting for CI to actually publish it.
So I have used WordPress for most of the time, but I never felt comfortable with it somehow. For one, it's often overkill for simple sites and blogs; second, it's not always easy to keep a WordPress site fast and secure. I have never considered alternatives like Medium because I want to own my content on my domain and be free to customize my blog as I please.
So I built DynaBlogger out of my own need and I've been hacking on it for some months now. With DynaBlogger I propose an alternative publishing platform that has all the essential features most people need, and nothing more, for a focused writing experience with no distractions.
DynaBlogger differs from a heavyweight CMS like WordPress in that it offers a simplified setup with no plugins and not too many settings etc.
However, it also differs from the several "minimalist" alternatives available because it's not too barebones and offers the ability to fully customize your theme with the built in code editor, if you wish. You can edit templates directly in the browser as well as upload assets and see changes in realtime. You can have multiple themes installed with one active, so when you make changes you can work on a copy and activate it when ready. No need to work on files locally, and upload a new version of the theme to see the changes with your content. Pretty handy. DynaBlogger doesn't have many themes yet but I will be adding more over time, and soon any user will be able to share a theme on a marketplace with other users (maybe even sell them at some point).
Perhaps the most similar blogging platform to DynaBlogger is Ghost, with a few key differences. DynaBlogger doesn't offer newsletters and memberships yet (perhaps in the future), but it offers easier customization of the themes and is more affordable. DynaBlogger also cares more for some details concerning for example SEO, redirects, publishing via email, etc.
It's a fully managed platform, so you can just focus on writing content and everything else is taken care of.
Tech stack for the curious:
- App written in Ruby on Rails
- Postgres for the data
- Redis for background jobs
- memcached for caching
- Hosted in a Kubernetes cluster managed by myself
I would appreciate if you could give it a try. You can sign up for a free 14 days trial and if you decide to stay, please use the coupon code 6649CB68 at checkout for a 15% discount. You can also import content from a WordPress or Ghost blog, so you can try it with your own content.
Let me know if you have any questions or any feedback, really!
If the frustration came mostly from bloated CMSs like Wordpress I would suggest taking a look at newer (and specially) headless platforms like Sanity.io
WordPress used to have a feature to edit PHP plugin code through the browser and it was a security nightmare. If you allow code editing, I suggest disabling it by default and when an attempt is made to enable it, send an email to the site admin to confirm before enabling.
Hi! The templates use Liquid, a very simple but powerful markup language based on HTML. It takes a few minutes to learn! It's the same markup used by Shopify - they created it https://shopify.github.io/liquid/
It's very safe though, it cannot execute harmful code :)
> Would you consider a pricing model to allow self hosting?
I remember I self-hosted a Ghost blog and mirrored/served the content with a CDN. It was tricky because once I changed a post, or altered the blog's code, the CDN had to be manually cache-busted if there was any change on the VPS instance of the blog. I done that a scary amount of times, making me believe blogs & blogposts are never 'finished'.
Some of my posts were picked up by various news outlets and that spiked the traffic. I was thankful I served it from a CDN because my Digital Ocean instance surely would have went offline due to the attention / hug of death.
Does Dynablogger handle traffic spikes gracefully?
Sidenote: unless you think you can get really, REALLY popular, you won't hit a CDN-deserving load.
My own homepage/blog is hosted on a cheap 100mbps box on OVH (the server has 16gb of RAM but the CPU is from ~10 years ago or so); I use self-hosted Ghost with a caching nginx reverse proxy. Cache expiration is around 30 minutes (I think). The price for such box is around 20 EUR/month for the whole server.
In order to crash such an instance I'd need an insane amount (of course now that I need the hard numbers I cannot find the log I had dumped at the time with Locust) of simultaneous users - something that never got close to happening, not even in those couple of times where I reached the frontpage on HN. My best posts got the box/bandwidth at around 1% the load that was needed to start dropping a significant number of requests.
> Sidenote: unless you think you can get really, REALLY popular, you won't hit a CDN-deserving load.
I don’t agree. A few months ago my blog post on CRDTs got 100k views within a week after landing on the top of HN. It’s just hosted on ghost with some poorly written custom theme that loads a ton of CSS for some reason. The VPS I host it from wouldn’t have dealt with the traffic - it served over 50 gigs within a week; and that’s over my VPS’s monthly cap iirc.
I have the whole site behind Cloudflare’s free plan and it handled the load flawlessly. And cloudflare also makes the webpage load faster for everyone. It seems like a no brainer to me.
100k hits over a week isn't much all things considered.
A static blog post that requires ~0,5 MB of transfer can be easily served from low powered server in an hour. That's only slightly above 100 Mbps and 27 rps. Servers are fast, much much faster than that.
Of course, poorly written code might make any solution too slow.
Agree, that's not much. Since I submitted this I saw up to 2K concurrent users at some point when I was 1st. Would be nice to have this sort of traffic for a week! :p
Yeah it’s not much, but I’d rather have cloudflare think about that than me. Especially if the only cost is 15 minutes of DNS configuration, it seems very cheap for the peace of mind.
Hi! The Kubernetes cluster is managed by myself so I have a pretty amount of resources for just 70e/month, and I can scale with one click. I plan on writing about the infra. Of course I'd like to think that this project will work, so I was happy to prepare for it. I enjoy devops besides coding so I am happy to do this. I need to consider what would happen if, say, a couple of my users hit the front page of HN at the same time or things like that. Already myself I have seen a peak of 2K concurrent users since I posted this, so I don't want to crash easily. :)
As for the CDN, it's also for performance/reduced latency.
Hi! It's hosted in Kubernetes so I can add nodes with one click if needed :) I am getting around 450-500 concurrent visitors at the moment and 2K total uniques since I posted and it's like nothing is happening. So in away this posting is a good test as I am still 1st on the front page I think :)
Static site generators I nice, but I just prefer hit and save to publish immediately. As for CDN and caching, DynaBlogger uses Cloudflare as CDN and does a lot of caching of the dynamc content too.
Hi! Thanks for your comment, I appreciate :) I am considering an option for self hosting. Need to think about it on how to make it work, but for now I am focussing on the hosted solution :)
One thing I'd like to see is your backup procedures. Is there any possibility that data could be lost 10 years from now? Also can I get a dump of my blog posts for my own safe keeping? (Especially since this is a project by just one or a few people, right?)
Other (probably more important things):
"About" seems to be the history of the project, but maybe it could be a high-level overview of the experience one might have using this blogging platform. Maybe a some screenshots to make clear how editing works and how close it is to the (more expensive) Wordpress experience.
The export feature is high on my list. It will be possible to export the content both as json and as plain static HTML that can be deployed as is on some static site host.
Backups: for the data, I have in place 3 levels of backups stored in two locations, including point in time recovery. Even in the worst case when something major happens say to the datacenter and I need to migrate somewhere else, I can do that with no to minimal data loss very quickly.
The assets (images etc) are stored in two object storage buckets with mirroring with two different providers.
Before launching I have been testing backups and disaster recovery throughly while I was testing various providers, so I have migrated the platform a number of times even with no downtime. Therefore I am pretty confident that it works :)
Thanks for the suggestions on the About page. I thought that people might be interested in knowing the "why" I built this given that there are some wildly popular options already. I will take a look at adding some info about the user experience. Thanks!
What frustrates me the most in every blogging platform out there is the requirement for every post to have a title. Inventing titles for every idea you want to share is immense pain and annihilates all the will to write. In fact I have never started (like I had an idea, some enthusiasm but postponed and forgot about them) with countless articles, computer programs, products and other kinds of projects just because inventing the title/name turned out too hard.
Sometimes it's easier to write your headline or post title after you've written your article. You can create a placeholder for a title (e.g. "Thoughts 001") and update it once the post is completed.
It has to start with a good abstract AKA TL/DR (which can also use formatting to highlight the key words) and/or end with a clear concise conclusion. Occasionally articles could fit in the size of an adequate abstract as a whole. This is what I would expect when I visit a particular blog directly, not a list of titles. And this would be enough, especially given the fact I visit a specific blog (I don't visit many) already means almost everything there is worth reading.
Hi! The app is written in Ruby on Rails :) So far I have around 100 users I think. I am getting new sign ups now since posting this so looking forward to seeing where this goes :)
Good job, functionality fits my needs and the UX is really good. One suggestion, seeing as though the whole thing is cloud-based and I can't really be sure dynablogger will be maintained forever, it'd be nice to have an option to export the blog content to your own machine. While I could just ctrl+c ctrl+v, I think it's more convenient to be able to export all at once.
Hi! Import and export is on my list already. It will be possible to export both in json format (so for example you can move content between blogs in DynaBlogger) and to plain HTML so in the worst case scenario, the user can just export to static pages and publish them quickly on Netlify or something.
Having said that, DynaBlogger doesn't cost me much to run now and it will be even easier once I have a bunch of paying customers, so I am definitely committed to keep it running for as long as I can :)
Nice Start. Just other day, I was thinking about a feature that is missing here, and see if you can add it, is the possibilities of the blog and site to be alive long after someone dies, probably to the end of the internet. Even after their credit card stops etc. ( Probably for a one time fee, add tech that is distributed etc ).
Nice work. I thought Ghost had this niche all tied up but goes to show there is room for innovation.
I personally love the devil I know too much: Wordpress, but if I was less techie, and frankly tightass (Wordpress hosting for unlimited sites is $60/y or so) I’d use something like this
Hi! DynaBlogger targets mostly users who don't need WordPress which is a beast and has a lot to offers because it's designed to create many kinds of websites, but like I said in the OP it's often overkill if you just want to publish content easily and quickly. It has just the features that most people truly need to publish content but it's not too "minimalist" at the same time. It's a balanced option IMO :) I know that there are several other options of course, but most of them are website builders (so kinda different) or by companies like Google who you pay with your privacy. So I am hoping that some people share the same views :)
pretty cool. The coolest idea I've seen recently with competition to substack is https://subclout.com/ where they almost cloned substack, down to the UI, but it runs on the clout system.
Congrats again on launching - appreciate the openness and the project seems interesting! Simple question here, as I somehow always seem to miss this feature on most blog sites, does this project handle pdf hosting as a blog post?
I would like to congratulate you on your value-prop, I think it is stated clearly and communicate to me what I can expect to get from this service. This, as you know is mostly not the case when people launch new things.
Hi, and thanks a lot for the kind words! I am "just" a developer so I have no experience with marketing and am not very confident about how I communicate what I propose, so I am very happy to read that you find it clear! Thanks again :)
Hola! Just checked and am not sure what you mean :) Pricing is $8/mo billed yearly ad $9/mo if monthly. Should I make this clearer on the home page somehow? Thanks!
> Pricing is $8/mo billed yearly ad $9/mo if monthly. Should I make this clearer on the home page somehow?
Yes. Regardless of how common it is or how many other companies do the same thing, "$X per month billed yearly" is a really stupid way to say "$12X per year".
Cut the bullshit. $96 per year. That's what the subscription costs.
Being mean certainly breaks guidelines, but I was not being mean.
There was a clearly stated argument, which was opinionated, and which you might agree or disagree with, but ad hominem remarks (like "you sound angry") are not a substitute for disagreeing or agreeing with substance.
Thanks! Very happy so far with how it's going. The app is written in Rails, and I use Postgres/Redis for the data and memcached for a lot of caching. Everything is hosted in a Kubernetes cluster than I manage entirely myself. Ask away if you would like to know more :)
For one it's a lot cheaper - I use Hetzner Cloud and I have plenty of resources for just 70e/mo. But I also like devops etc so I don't mind it :)
Later of course I can switch easily to a managed service if needed.
I dunno. What makes me wonder is, if Hetzner Cloud can offer an excellent service (great performance, very reliable, very good control panel etc) for those prices, try and imagine how much money AWS/GCP/Azure make!!
Luke W and I made a blog platform for him in an afternoon on a Saturday years ago. Works pretty well. They aren't hard to build IMHO and if you write it yourself it just does exactly what you want it to do:
For me witb DynaBlogger the thing that took most of the time was implementing the themes so that they can be edited by the users. A blogging platform isn't rocket science but can be challenging for some things. :)
Hi, great implementation, i was curious if there was any plan to offer a pricing per blog or something of the sort. I don’t write too much even though id like to, so 8$/mo seems steep to justify me writing once in a while. So a pricing per post or something might be awesome
Hi! I haven't considered pricing per the amount of content or similar structure, I'm afraid it would become too complicated as pricing. What do you suggest? What kind of plans? At the moment I have just one plan with pricing per site either monthly or yearly to keep things super simple.
I could introduce a cheaper plan perhaps. What kind of plan would you be interested in? Thanks
Hey, so as i mentioned i dont write nearly as much content as id hope. So say for example if the pricing was more like 1-5 $ / post for first n posts in the month and then less $ for more than n posts in the month. etc that way, for people who are like me and publish less content, it is still feasible to subscribe rather than paying flat 8$/mo regardless of if i publish or not. For people who publish way more often, it can still make sense since the overall costs will be similar.
Only slight downside i can see is that it deters the person to publish content since it costs to do so, but i think if you are driven enough to publish anyway, a minor fee is not something you'd mind, esp if you are choosing this over some of the free platforms.
I need to think about it because there is a threshold in terms of price below which it's no longer convenient for me due to the payment processing fees. Charging per the number of posts means that if someone publishes say only one or two posts in a month, the transaction would be too small and there is a fee per transaction so it might not be convenient. Will need to think about it. Thanks for the suggestion!
hi! I think that blogs won't "die" easily. Sure social media is all the rage these days but most quality content is usually found on blogs in my experience. I find help in blogs on countless occasions in my job and I am sure that this is true for many other people too :)
I really agree. For example in infosec, lots of valuable texts are published on blogs. Then the author spread them via e.g. Twitter. I don’t see that way of publishing going away.
Exactly! I find it funny that every now and then someone says that blogging is dead and things like that. I don't see it happening any time soon as well as I don't see people publishing e.g. long form essays etc on Facebook.
Hi! You're the second person who asks about this :)
For now I am focussing on the hosted version but I will consider a self hosted option at some point.
I like this phrasing. Many of the best things are made out of frustration - it might be one of the most important building materials for software.
‘Cool blogging platform! What’s it built out of?’
‘Frustration’
Definitely better than building one out of fear, anger or greed.
Actually, in some seriousness, I think stackoverflow was built out of anger - and used that to clearly define what their differentiators and value proposition were. So maybe hanging on to that original motivation could be an important part of setting out what makes this different.
Anyway, congratulations on launching.