I'm using hugo, not jekyll. But I don't think it matters which site generator you pick. The key point is using something that is code driven. And then have AI drive the code changes. Basically all routine site maintenance and updates is now controlled via agentic coding.
We use guard rails and skills to impose structure and process. This includes tone checks (and fixing), making sure audio transcriptions are in sync with articles, ensuring everything is tagged correctly, dealing with translations and approved lists of translations of key phrases, SEO checks and much more. I've been dialing in a lot of this in steps. You can start without most of this. But essentially a lot of manual work melts away when you get a bit structured on this. Like the article, we also use vector search embeddings. Our search actually uses the same model and runs it in the browser via web GPU. I also use it for related articles.
Also we've been experimenting with using reveal.js for presentations. Same principle. Forget things like Keynote, Canva, etc. Reveal.js is meant for programmers. But if you replace those with agents, non technical people can prompt together some really amazing decks. Replacing applications and UIs with code driven systems removes the need for those applications and UIs. And using AI to drive those code based systems removes the need for having developers in the loop. Our non programmer CEO who was a heavy Canva user is now doing decks and huge website updates this way now. Pretty scary actually. I don't think he'll use Canva again. I'm barely involved beyond setting up some basic plumbing. One party trick he likes is adapting decks to customers by integrating their house colors and visuals. Only takes pointing the AI at their website.
https://querylight.tryformation.com/ is a hugo demo site for the search capabilities. It hosts the documentation for the vector (and lexical) search library I use on our websites. The entire documentation site is managed as I describe above.
I went the other direction: instead of replacing the CMS, I open sourced a FUSE like layer [1] that mounts any backoffice system as a filesystem while the source of truth is in WordPress, Mysql, Postgres, .... Most backoffice systems map naturally to a file tree, so you mount them locally and let your agent read and write through that.
In the POC I ran, I mapped a whole wordpress site as a fs which was versioned controlled via git so when you did a git checkout of something, the whole db would get updated on the fly
From a glance at you documentation I take it that I could just write a plugin for filestash ?
The blog is running already on hugo on full llm automation but I had though it would not work for documentation (this is for non-techies so want something more product-manual -like and less SDK-docs flavour) or landing page that well.
So this is you company’s site and it’s on Hugo? https://formationxyz.com/
Interesting facts:
- the domain was registered a few weeks ago
- most of the heavy lifting was done by our non technical CEO by prompting codex; not by me.
- he got a bit carried away with some features and he was able to pull off a little robot (powered by vector search and embeddings), audio transcriptions, and a few other nice features. He has a product/ux background and a good eye for detail but no coding skills whatsoever.
- we use a lot of skills and guard rails to guide content generation, SEO optimization, etc. Our SEO agent does competitive analysis on a schedule, figures out optimal SEO phrases and maintains a list of approved SEO language.
- our content generation skills guard against typical AI slop smells, weaves in SEO language where possible, and uses a sub agent to act as harsh critic on content. AI slop only happens if you let it happen. You can see on the querylight documentation site that I have a bit more of that there. I need to improve the skills for that one.
If you want to get your feet wet with this, I would just recommend doing it and start with simple changes. Use Claude Code, Codex, or whatever you prefer.
One of the first successes I had with another website was "add the logo for company x to the logo wall". It went off and figured out the website, got the logo svg, figured out where to put it and hooked it up in the right place. For me that was a oh "it can do that now" kind of moment. A lot of content changes are like that.
In my view, a CMS is intended for people doing stuff. If you transition that stuff to AI Agents, why keep the CMS around? And if AI does all/most of the coding, it's not such a big leap for non technical people to get their hands dirty anymore.
I wrote about my journey from WordPress to Jekyll at https://brajeshwar.com/2021/brajeshwar.com-2021/
I moved to Hugo after years of suffering long builds. Jekyll's stack complexity meant CI/CD was hard to implement and it just wasn't that fast. Hugo wasn't just fast, it had everything baked in. The only problem (back then) was the lack of useful feedback. It'd just build omitting half the site.
Now I use Astro. Simple to understand, none of the dark magicks that Hugo relies on to work, and Quick Enough™. It's well supported and it integrates with all the frontend/design tools I was already using.
Pagefind is stupidly easy to integrate. And it's what you probably ought to be using whatever your SSG. https://github.com/shishkin/astro-pagefind
I won’t argue with their reasons to move (which don’t stack up for me either but agree to disagree).
I was all for them until trying one, now I don't understand.
If all you need is a, well, static website without forms or such, Hugo is an easy winner.
It's a pain in the ass to use.
With SSGs, you have a few options for comments, like Disqus, but the ad-free version costs money, it's slower because it needs to load JS, and your comments are owned by a third party company. Contacts forms can be built by integrating an external API. And anything else that requires storing data will require an integration with a third party service of some sort.
SSGs are a great concept but they're mostly for nerds who get boners by seeing over-engineered systems. They're also great for companies like Cloudflare because they can sell you services that come for free with Wordpress (CMS, image uploads, databases, workers, etc). For serious blogging, I'd opt for Wordpress.
No server side things to worry about. It's super clean. Jekyll, css, js, GitHub and cloudflare is such a clean and refreshing setup.
Multiply that by every nonprofit without tech knowledge and that's a lot of potential Jekyll fans.
But yeah, a bunch of html pages is over-engineering.
PS: you don't need to use Disqus if you don't want to, there are a myriad of free open-source alternatives you can self-host that will take a fraction of the maintenance work of self-hosting a Wordpress instance.
Social accounts aren't necessary either. Users can simply not login, or use their email address. If you still want social accounts, it's a one click plugin install. With SSGs, you have to bake it yourself or buy a comment system that includes social auth.
SEO is a big deal, and Wordpress doesn't allow it by default. However, you can install a minimalistic SEO plugin like "The SEO Framework". If you want more advanced SEO, you can install a plugin like Yoast. With an SSG, you have to build your own theme that will allow SEO using frontmatter from your HTML pages, or use an existing theme that has that functionality.
Adding plugins introduce additional attack surfaces, but if you're building a full-fledged SSG to rival wordpress, you're connecting to various APIs, and writing your own code that could easily introduce security bugs. Wordpress plugins can auto update, but SSGs require you to fix bugs on your own time.
Wordpress version changes can break your site, but you can easily backup and restore your content for easier migrations these days. SSGs can also break your site if you want to keep the generator up to date. For example, if you use Jekyll 2, you'll have to spend some time refactoring your project when upgrading to version 4. Now, if you have integrations that you built yourself, this could take considerable effort.
I don't personally see how an SSG can compare to Wordpress. Sure, if you enjoy writing Markdown files in Vim then go ahead, but for people who mainly want to write content and provide a service that visitors can interact with, Wordpress wins all the time.
Have a look at https://service.polymech.info/user/cgo/pages/poolypress-cms, agentic CMS, translates, creates and manages articles with a few prompts, widget aware.
But these days any new site I build is on NextJS since coding agents make it a breeze.
You can actually still create new Pages projects, but it is de-emphasized in the UI in favor of Workers.
We've done a lot of work to make static sites on Workers just as easy to configure as Pages was. Have you tried it lately? Would love to hear what aspects you feel are still more complicated than Pages.
Just to be clear, we will not break existing sites using Pages. We will either auto-migrate them to Workers once we have all the tools in place to do so, or we'll keep supporting Pages forever.
I also don't want to tie my site to disqus or other 3rd party cloud services and their implication on GDPR.
And running this opens you up to security issues you were trying to avoid by going the SSG route. In a way you could just keep WP, then.
(Also, WP has this beautiful ActivityPub plugin that makes your blog a fediverse account that people can subscribe to and even comment on your posts from Mastodon/Pleroma/etc.)
> legacy WP blog ... users should be able to ...
(What's your blog about?)
It was the last thing using MySQL, PHP, and Wordpress on my site. 3 big things to not have to keep up-to-date and secured. I can check in markdown to my repo, it builds the site, and Nginx serves it. So fast, and secure.
There is a WordPress migrate-tool that worked flawlessly.
I mean, it’s tempting though.
