…there were plants.

Saturday morning, Home Depot. Nine herbs, two seed packets, one bag of potting mix, and a small pile of plastic pots. By Saturday afternoon, everything was planted, watered, and arranged on the porch. That night the hardware for what is to come was ordered and now I impatiently count the days. (Or begin writing about the process)

Occasionally, I will have a thought or an idea come across my mind, I get stoked and begin scribbling notes and plans as to how I am going to proceed and what it is going to involve. I gravitate towards the possible and forecast the best version of what it can be before narrowing down into what a minimum viable solution would be. I would say more often than not, those ideas end up shot down due to either the idea maturing from conceptual thought into something unrealistic or simply deem the juice not worth the squeeze.

But Tuesday May 5th was different (for a number of reasons really). Most of you reading were probably working that day, but as a recently laid off Designer I have spent the last few months building this site, collecting my previous work, running, fishing, training a puppy and so on. To build this site I have dove headfirst into building my AI workflow unburdened by the guardrails of a paid role. After the site was done, the fire to continue learning was still burning bright. I had been looking for a new endeavor and luckily for me it was sitting on my back porch. Normally I am there watering my Trident Maple Bonsai, but on Tuesday morning I looked at my girlfriend’s dead herbs and a thought passed my mind that led me to this project you are now skimming through reading about.

I will admit, this idea was the culmination of a few things..the people that surround me and the content I consume, my brother recently shared with me a job opening at ATOMSlink. They have an inspiring body of work ranging from automating kitchen lines to haulage systems. Not that garden herbs and industrial system share much in common, but there was enough to inspire the thought to overcomplicate the peaceful process of tending to ones herb garden.

What better way to embrace the future of automation with the natural process that I admire everyday. I am not sure the exact term for this, but I’ve found that Solarpunk or Permacomputing come close. So if I wanted to be fancy it would be one of those.

The Vision

We will get to the actual v0 details, but first let me paint the picture as to what I am aiming for. An outdoor cabinet housing water reservoirs flanked on either side by metal shelving that hold 8-10 herbs each. Solar-powered ESP32 controllers handle watering autonomously. Soil moisture sensors per pot inform watering decisions. Pi 4B indoors runs the dashboard, database, weather integration, and aggregates data from all edge devices over WiFi via MQTT. Not to mention a Web app accessible from any device on my home WiFi shows live status, history, plant journal and insights. The whole system fits in a few boxes when I move.

This is less of a gardening project and more of a proving ground for myself . Sharpening hands-on skills across IoT, embedded systems, backend infrastructure, and the full-stack thinking required to make hardware and software work as one system. The garden is small and low stakes which makes it the right place to learn patterns I might extend further later on.

The plant roster

The nine plants spread across three sun zones on the north-facing porch:

• Sunny railing zone: rosemary (in a 12” glazed ceramic I already had), lavender(my girlfriend’s favorite), thyme (in a 6” ceramic-coated pot I already had), oregano, sage
• Mid-deck partial sun: garlic chives, parsley
• Near-wall shade: spearmint (I learned that mint can dominate anything its next to, so it is isolated), cat grass(shout out Cheeto)

I may add a 10th as my Trident Maple but I want to save some me time out there.

Collecting and Planting

I’ll admit I don’t know much about gardening, but that wasn’t going to stop me. I have momentum and I am intent on capturing it.

Near mistake: Garden soil and amendments aren’t potting mix. The bag I almost bought said “Garden Soil for Flowers & Vegetables” on the front, but the icons in the corner said “Clay Soil Conditioner.” Catching this before checkout saved a season’s worth of root rot.

Correct choice: What you actually want is a bag explicitly labeled “Potting Mix” with the Mulch & Soil Council “Premium Potting Soil” certification. The Kellogg Raised Bed & Potting Mix met both criteria and had the right ingredients (recycled forest products, coir, perlite, compost, peat moss).

For my Mediterranean herbs (rosemary, lavender, thyme, oregano, sage), I mixed in 30 percent-ish extra perlite for sharper drainage. Those plants prefer drying out between waterings; the regular mix would hold too much moisture.

There is not much to share here, I’ll have to get photos next time to share the process. But they ended up in their pots with proper drainage matter on the bottom. I had some small rocks handy, but one thing I learned the hard way with one of my previous Bonsais is to mind the bottom of your pots so they can drain properly.

Hardware ordered

After the dirty part was over I moved to my favorite hobby, over researching any purchase I make. I could belabor the point here but this is already getting long so I’ll cut to it. I biased towards over compensation in the Raspberry Pi, I probably don’t need 4GB but what if later on I do? So I went for it.

The brain of the system is on its way:

• Raspberry Pi 4B 4GB (Vilros) — the orchestrator
• Vilros clear case with built-in fan
• 2× ESP32 Type-C dev boards
• SanDisk Max Endurance 32GB microSD (the durability tier matters when the Pi will be writing to disk constantly)
• Official Raspberry Pi 15W USB-C power supply
• Micro HDMI to HDMI adapter
• USB-A to USB-C cable for ESP32 programming

Total V0 investment: $309.49 (plants + soil + hardware).

The full plan (as of right now)

V0 (Complete): Plants are in pots. I’m watering by hand and paying attention to which ones dry out fast, which sit damp, when the sun hits the railing. Whatever I learn this month will tell me how to set the system up later.

V1 - Proving the Architecture: Hardware arrives. The Pi gets set up indoors, the ESP32s get programmed, and I’ll wire up a demo of sorts while I’m still figuring out the firmware. The whole point of this stage isn’t to water plants — it’s to prove the system can talk to itself end to end. Click a button on my laptop and confirm the unit responds. Once that works, I will hook up the valves, pumps, solar panel(s), and the system should be ready to start watering the herbs on a schedule.

V2 - Implement the System: This is where it gets fun. I’ll put the Soil sensors in so the system starts watering based on what the soil actually needs instead of what I told it to do at 3pm. The dashboard becomes something I’d actually want to look at. The reservoir, pump, and electronics get hidden inside a cabinet, with plants stacked on shelves on either side. From the porch, it’ll just look like a planter setup. The technology ideally disappears.

V+ - Expansion: Whatever this becomes once it’s actually working. Probably a second unit. Maybe a camera. A weather mointoring aspecct has me thinking and it might end up its own project that feeds into this one. Whatever’s interesting at that point.