Maker Journal March 2026
March always feels like the real start of the year for me.
After the whole Lunar New Year period, everything kind of resets. You stop drifting, start tidying things up, and mentally switch back into “let’s go again” mode.
So I kick-started my product-building engine again.
With AI in the mix now, things definitely move faster. But at the same time, it opens up many new paths, which means more things to learn, more things to try, and, honestly, more things to get distracted by.
RebuildHQ
This one went through two revamps in a pretty short time.
Initially, I tried going full AI for design — just prompting and letting it generate everything. Turns out not great 😅 . Something always feels off when there’s no real design direction behind it.
Then I switched approach. Used Google Stitch to guide the design, and let AI handle the generation part. This round feels much better.
Also upgraded the underlying AI model, hoping the output becomes more meaningful instead of “looks smart but actually meh”. Still early, but I like where this is heading.
MD2Page
MD2Page got a refresh as well. Surprisingly, this time, AI actually nailed the landing page. Didn’t need much tweaking, which rarely happens.
The bigger change is the pivot. Originally, it was about generating landing pages from markdown.
Now it’s more about displaying markdown nicely as a page. Subtle difference, but feels more usable. Not sure if this direction will work yet, we’ll see how it goes once I start pushing it out.
On the backend, I experimented with both OpenAI skills and Anthropic skills.
Honestly, it got too complex, and token usage went a bit wild. So I went back to a simpler setup using just the OpenAI API. Sometimes simpler is just better.
I’m also thinking whether to explore MCP or a CLI approach for this, but probably after I get some real feedback first.
Statify
Finally shipped Statify during my Product Weekend. I had some rough ideas and prototypes last year, but nothing really solid.
This round, I changed direction:
- Package it as an open-source tool
- Turn Filament widgets into APIs
- Build a Chrome extension as the first client
So now:
- Backend → exposes widget data via API
- Extension → consumes and displays it in the browser
Both are live now. Feels good to finally make this more “real” instead of just sitting as an idea. The next step is obvious — marketing.
(Which I always delay, but yeah, time to face it.)
ClipSend
Didn’t do much coding here this month. Most of the time went into thinking.
I still really like the core idea, but something feels off with the positioning and growth. One idea that came up is the satellite product ClipSend.
Feels like a more realistic way to grow it, rather than pushing the core product directly. Still very early, I’ll probably explore this further (with ChatGPT 😄) and see if it’s worth committing to.
Random Thoughts (and distractions)
Because Codex and Claude both are giving extra usage credits this month, I’ve been fully utilising them. Not sure how things will feel once that’s gone in April 😅
Also started playing more with Veo 3 videos. It’s fun, but also another rabbit hole.
Feels like everything is moving, but also hard to focus on one thing.
Wrap up
Overall, March was a good month.
I shipped quite a bit, explored a lot, and got back into the rhythm after the slow start of the year. Not everything is clear yet, especially on direction and focus, but I’m okay with that.
Momentum is there. Now the goal is simple — carry this into April and don’t drop it halfway.