Weekend Hack: Making My Blog AI-Searchable (No Flames Required) 🔥🚫

First of all—thank you. 🙏

For real. This community has been nothing but encouraging, and it was a huge surprise to end up in the top seven last week. I didn’t expect it, but it’s appreciated more than you know. 🫶

---

TL;DR

If you're looking for a quick and efficient way to increase traffic to your DevTo blog posts (or anywhere else with a little tweaking)? Try my 10-minute DevTo crawly mirror. Here's how it works:

1. Clone the repo. Enter your DevTO username as an env variable in the repo. Delete the existing gh-pages branch.
2. Scheduled GHA Workflow off-peak or manually trigger.
3. Static HTML hosted via GitHub Pages
4. Canonical back to Dev

---

Starting With Nothing 🔦

So, this entire thing started with a conversation.

🦄 About what? I honestly can’t remember now—Copilot, debugging, something in that ballpark. Doesn’t matter.

But somehow—like so many of my rabbit holes—it mutated and morphed until I wasn't exactly at the same place I had started. I really can't explain the series of events that happened after that. I blame it on curiosity more than anything, but I did the one thing I swore I'd never do: I asked ChatGPT to do a quick search—pull one of my posts.

And... nothing. Like the post didn’t exist. 🕳️

Okay. 😅 Fine. Let’s take a step back. Next, I went straight to Google and typed the same thing. Results were better—sure. Scroll, scroll, scroll... and there it was. My post. Not at the top, but alive. Win. 😏

🦄 Real talk: Do I care about being #1? No, not in the slightest. Honestly, I love my invisibility bubble—keeps things peaceful in my corner of the internet.

But here’s the thing: invisibility and AI don’t mix. Because when I need an answer, I don’t open a browser anymore. I yell at my phone: “Hey Google!” Or I pop into ChatGPT or Gemini.

If my posts don’t show up there, they might as well not even exist. 🫤

---

Déjà Vu 🔮

So I’m now blessed with the knowledge that my posts are basically hiding out in the Room of Requirement—you know it’s there, but only if the stars align. And that makes it hard to share the advice I keep repeating everywhere else that will listen:

• “Yes, you can use it for debugging.”
• “Of course it can generate the tests.”
• “Docs? Oh my god, PLEASE let it write the docs.”

By the time I’ve said it for the fifth time, it feels like I’m stuck in reruns.

🦄 Honestly, some days I feel like I need a theme song: “Previously, on AI Adventures with Ashley...” 🎵🎶

Sure. Sometimes, I let AI run wild (it's more fun that way!). My Coding Agent kills it with documentation. I even built a chat mode that plays architect—reverse engineers a repo and spits out mermaid diagrams like it’s dealing cards in Vegas. That one runs in GitHub Actions, then opens a PR, and I'll merge once everything checks out.

But the everyday stuff? It’s not dramatic. It’s small instructions, nudges, prompts. Rinse, repeat, coffee, repeat again. ☕️ It's slow, incremental improvements to be just a little better than yesterday. 🤏

And what drives me absolutely up the wall? The extremes. Either “AI will replace every dev tomorrow” or “AI is useless, don’t bother.”

🦄 Honestly, it’s like sitting court-side at Quidditch—the little guy zips across the field so fast you think you’ve finally caught up, but nope. He’s already off to something else... and it’s almost always completely wrong, too. 😅

---

Writing It Down 📕

So my solution: write it down. That’s what most of these blog posts are. Whatever comes up that week—whether it’s me digging, someone else asking, or just another repeated question—if I haven’t covered it already, give me a week. I'll get there.

The beauty of this system? Instead of explaining everything for the tenth time, I can just say: “Glad you’re trying this. I think my blog might help—here’s the link. Let me know if you’ve got questions.”

See. Simple, right? 🙂‍↕️ But even that wasn’t enough, because usually by the time I arrived on scene things were in somewhat of a state already. Mostly because what AI was helping with had a knowledge cut-off of January and everything else is wading through Microsoft or GitHub docs looking for any little nibble.

So I set out to find a way to make my DevTo posts truly crawly—without Dante's DevOps nightmares playing on repeat at the top of every hour.

🦄 Sure, I briefly considered the real solution—custom website, full hosting stack, the whole front-end shebang. But I'm a backend engineer for a reason! My personal site is designed for APIs not MUI. I'd really prefer to keep it that way!

---

The Hack in Action 💻

So, I did what any other respectable developer would do with an extra 15 minutes and spark of an idea: I hacked it. What do you get with 3 hours, a solid round of "will OCD win again?" and ChatGPT + GitHub Copilot tag team special? Exactly zero upkeep—mission accomplished. ✅

GHA cron → Dev API → static HTML → GitHub Pages → robots.txt

That’s it. No CSS. No fancy layout. Just static content, scheduled on a daily push. Best of all? ChatGPT could finally find my posts. Gemini too.

🦄 True story: I spent longer arguing with Leonardo to make a banner image than I did setting up the whole mirror.

---

Add A Little Sparkle ✨

I'm in no way an expert when it comes to SEO or even scraping in general. I did happen across this tidbit in the 15 minutes of research I did on the subject. I'm still testing this to see what the impact looks like, but seems promising. 😇

In your DevTo settings, you have the option to use the current rich-text markdown editor we all know and love or, you can switch to the legacy version. It's kind of an all or nothing deal, as far as I can tell. So you either have it everywhere or nowhere. But, it does give you access to the yaml front-matter for your posts and the ability to define your own brief description instead of relying on the first ~160 words you post.

It seems like an awfully lot of work though, and I've kinda gotten used to the new markdown buttons (that I never use). So my solution? Let's just hit buttons until I find one that works. 🤓

---

What's Mine is Yours 🍀

Whether you’re after more profile views or just tired of repeating yourself like me, this trick works. You can have my quick and easy DevTO-Mirror running in less than 10 minutes. Just copy it, fork it, clone it, star it—whatever you like. And please, avoid Dante’s DevOps while you’re at it. 🔥😈

---

This post is now searchable 🛡️

Responsibly indexed by me & the robots. Fueled by coffee ☕️, steered by OCD, and completely bypassed the 9th circle of hell. 🔥🚫

---

🦄 No, I'm still not finished getting this character exactly where I want her to be for the banners. But they're so close that I had to use it. Minor tweaks still and an AI character (or several) to tag along still to come 💖

Comments

[Ingo Steinke, web developer]

And what drives me absolutely up the wall? The extremes. Either “AI will replace every dev tomorrow” or “AI is useless, don’t bother.”

You have a point, but what's even worse, 99% of AI-related discussions view AI tools from a user's perspective, taking for granted that AI is cheap or free. Reports and discussions about underpaid AI trainers, copyright issues, and environmental aspects like energy and water consumption are marginalized and drowned below the 1000th AI is good or AI is bad post. Companies like Geforce, GitHub and OpenAI got us, as developers, exactly where they want to, and we'll start paying the price when there is no easy way back to pre-AI days.

[Ashley Childress]

I’m not usually writing about “AI theory” — I’m writing as the user. Because I am the user. And you’re going to have a hard time convincing me there’s some shadowy master plan at work here. If “exactly where they want us” means we’re using the tech that finally broke out of the labs and landed in public hands? Then sign me up. Please and thank you. 🐇✨

As for “pre-AI days,” which ones do you mean? The 50s before the idea existed, the 80s before anyone could scale it, or five minutes ago before it went public? None of those are real exits. Either you erase Turing from history, or the tech still shows up eventually. Worst case, it’s locked up for the 1% — and if that’s the alternative, I’ll take the current mess over that any day.

Now, on your first point: underpaid trainers. I just don’t buy it as an “AI problem.” Overworked, underpaid people exist in every corner of the world — tech included (and myself some weeks) — and that’s not new or unique to AI. That’s work-life balance collapsing. Employers absolutely need to do better, but employees also carry weight here: every time someone just shrugs, stays quiet, and powers through, it makes it harder for the rest of us to fight for change.

Renewable energy? Yes. Copyright? Definitely — though I’d love to see actual experts in the room instead of another parade of lawyers pretending they’ve read a whitepaper. Not that I’m holding my breath. But still, the bigger picture is: we’re in the awkward beginning stages, with tech outpacing everything else. To me? That’s exciting.

So yeah, there are growing pains — and I’ll take them. Because I’m not getting in some time machine just to rewind into a world that’s slower, smaller, or duller. The future’s the only direction I’m interested in. And if that’s the trap, I’m bringing the snacks. 🚀🍿

[Ingo Steinke, web developer]

Thanks for your elaborated and thoughtful detail. I see and I now totally agree on most of your points:

Worst case, it’s locked up for the 1% ... Overworked, underpaid people exist ... not new or unique to AI ... Renewable energy? Yes. Copyright? Definitely — though I’d love to see actual experts in the room instead of another parade of lawyers pretending they’ve read a whitepaper.