n8n is pure awesomeness

Source: Dev.to

One of the big benefits of having a Homelab and plenty of services running locally is that I can start connecting them together and using them for various things. TL;DR, automating like crazy. Not to mention that there are some things running on the wild internet that I do not plan to self‑host; they’re still valuable to me and I want to integrate them into my whole ecosystem.

I do want to point out that you can also use some of these services by paying a subscription—as most things nowadays work. You could use free tiers, but once you hit a limit you’ll end up paying.

Whenever I do automation at home, 90 % of the time it follows a very repetitive pattern:

1. Put together a bunch of Bash scripts that call something to retrieve data.
2. Format or parse the data and enrich it as needed.
3. Send a payload elsewhere and eventually achieve the automation’s objective.
4. Profit.

That approach worked for a couple of years, especially for notifications whenever something went down in the Homelab and Grafana couldn’t handle it. As more services ran locally and more integrations appeared, I discovered no‑code automation tools that made things so simple that even someone who isn’t a “tech bro” could use them.

On top of that, the hype around AI made automation even more desirable, because it can be expanded into “AI Agents.” To me, those are basically (Hot Take of the week 🔥) big cronjobs that either run on a schedule or react to an event, following an event‑driven architecture. They still use the same methodology described above.

What am I using and what am I doing with it?

I’m running a local version of n8n. You can start with their free tier, but since I have the hardware I always try to self‑host to keep my data under control.

Get the weather

This one is kind of dumb, but as a busy father I need to be prepared for the weather so the kids wear the proper wardrobe for daily activities. I designed a simple workflow that runs every day at 7 AM and sends me the current weather via Telegram.

Daily review of my Todoist tasks

Todoist is my default task manager. As someone who tries to follow GTD, I dump everything out of my brain into the tool quickly and review the list of tasks now and then. Previously I had to open each project to find overdue items without labels or dates. No more—my workflow kicks off, filters tasks, and adds a @review label to anything that doesn’t match my criteria, making it easy to spot and organize.

Sync Tasks‑Issues in GitHub

Some of my Todoist tasks are related to small projects on GitHub (e.g., my Twitch‑bot or k8s Homelab applications). When I want to try a new tool I create a “ticket” for myself. Having the task only in Todoist is fine, but the projects are public, and I once had issues open in a repo for a Neovim plugin that I ignored for weeks because I rarely check that section of GitHub.

The workflow now syncs Todoist tasks with GitHub issues. It also uses a small database to keep track of what has already been created in either system, preventing duplicates.

Adopt a Dog

This is the one I’m most proud of. I created a web‑scraper that checks my local shelter’s site for new pups, extracts details (breed, size, date of birth, description, etc.), stores them in a table, and sends me a notification via Gotify—another service I’m using for all things notification‑related.

Looking ahead: AI integration

That’s only the beginning. AI is here to stay, so the next workflow I build will use a model to do something fancy. I haven’t found the perfect use‑case yet, but I’ll be ready with the infrastructure to run it—thanks to Ollama, I can also run my own models locally.

Conclusion

What I love about the whole n8n ecosystem is that it can be triggered by a lot of different inputs or events. Most of the ideas I’ve come up with are still cron timers, BUT I’ve been playing with webhooks, and getting them to work and trigger something took me less than three minutes. This is great because n8n abstracts the complexity of standing up a web server and listener for you; it can dedicate a specific endpoint for each workflow, which can then call other workflows.

I’ve seen YouTube videos of people building automations with 40 + nodes that perform convoluted actions, and others claiming that these so‑called “AI Agents” are being used to make real money. I don’t doubt it—n8n can integrate with thousands of other systems, so for businesses this makes perfect sense.

I would recommend that you try it out and get some automations going. You don’t need to be a proficient coder to get these working, so the barrier to entry is incredibly low—and that’s a good thing!

The days of writing bash scripts are over for me…?

Until the next one,

Adios 👋