PersonalTab

Save notes, make lists, build launchpads, monitor sites via RSS. 3rd mini-app built in-house with AI and no programming experience.

I have stumbled way more than I would have anticipated, but here it is: my third mini-app created with AI assistance is ready.

PersonalTab is a productivity Chrome extension that I created to challenge myself in building something useful with AI (and no programming experience) that supports some small and typical personal productivity needs.

What Does PersonalTab Do

PersonalTab organizes notes, info and resources onto a customizable private dashboard that is displayed every time you open a new browser tab.

Inside this dashboard you have by default four widgets: a Notes widget, a Links widget and a List widget…

a) Notes
write notes, save text snippets, codes, or other info that you may need from time to time. Each Note widget is a small text editor ready to capture and preserve anything you write into it.

b) Lists
create lists of things to do, items you need, tasks you need to remember and carry out. Each new widget created allows you to list tasks, items, places and to easily re-arrange and mark them.

c) Links
build link launch-pads from which to open websites and apps you frequently need to access.

d) RSS Feed
monitor latest news and posts from specific websites, newsletters, social media channels. Just create a new RSS widget and paste the RSS feed of the channel you want to monitor.

You can create more of these widgets when you need them by simply clicking the plus sign appearing on the top right of the PersonalTab page.

Other key features:

• Extensible
  It is possible to create as many widgets as desired and to order and organize them on the page as you wish.

• Customizable
  Each widget can be sized, positioned and titled according to your needs. Even the PersonalTab title can be edited and customized while items inside Lists and Links widget can be re-arranged in any order.

• Local
  Everything is local. Nothing is uploaded or exported to the cloud.

• Safe

  You can backup all of the contents of your PersonalTab and restore them anytime on the same computer or on a different one.

• Instant

  PersonalTab is always one-click away from any page or app you’re in. It requires no login, no passwords, no two-form authentication or passkeys. It’s just there, always ready for you just by opening a new tab.

Here is a short (4:47 mins) video introduction to PersonalTab.

Why I Built PersonalTab

I have built PersonalTab around my need for having a private, safe, local space, where I can write, organize and keep notes, lists and links that I often need without having to open different applications, emails and hard-to-remember documents and files.

PersonalTab is always there, instantly accessible and super-easy to use. That’s what I like about it the most.

I also wanted to have a customizable space where I could monitor in real-time specific news feeds, newsletters and websites.

For this purpose I have integrated inside it a RSS feed reader, that allows me to create and monitor as many RSS feeds as I need.

The Story of PersonalTab

I was initially inspired to build PersonalTab when in the summer of 2020 I discovered through ProductHunt a little-known Chrome Extension called Dwij NewTab (not available anymore).

It had a nice minimalist interface, felt a bit like Trello, it had some interesting little touches and a slick UI that I wanted to improve upon.

So I didn’t think about it too much, I hired a trusted Indian programmer and provided him with both the inspiring existing app and with a brief containing my new interprepation of it and all the changes and features I wanted to add.

Within three months (60-90 hours of coding, revisions and fixes / apx $1,000 in costs) I had a fully working version ready to be used.

I installed it, used it and have not removed it from my computer ever since.

But I never released to the public.

Maybe it was because at that time I moved from Terceira island (Portugal) to Holbox in Mexico, starting a completely different life, or maybe because I still wanted to add a few refinements and little frills, but fact is, that I gradually forgot to official launch and share this project with my readers.

Time passed and I became accustomed to having it at my disposal every time I opened a new Chrome browser tab.

Three months ago, as I pushed myself into this monthly challenge of creating and releasing a new mini-app monthly with AI assistance, I started listing what needs and apps to go after for this new project.

That’s when I realized PersonalTab was still sitting there, and thought of rebuilding it from scratch.

And so I have done.

Between May 15th and June 20th 2025 I tried several AI coding environments and tried to rebuild from scratch PersonalTab.

I instructed AI on what I wanted and started building. I bumped into so many issues and problems I can’t even recall them. Worst of all I have had to restart the work many times and with different tools.

I gradually learned many tricks and rapidly adapted to AI demand and way of proceeding.

When I got a few inches from the end, with the many little features I had dreamed of integrating, the code exploded and I found no way to proceed from there.

I suppose that the code got too big or intricate and the AI itself was not able to reliably handle it anymore.

Thus, the final app I am releasing now, doesn’t have a few little fine touches I wanted, but its core features and abilities are all there.

The absurd novelty is that I can’t refine it or add to it. At this stage, anytime I try to add or fix something, the code just explodes and AI vibes out a new app version that doesn’t remotely resemble what we were building together up until a minute before.

So, to have a next version of this app, it looks like I will have to rebuild PersonalTab, once again, from zero.

What I Learned

Vibe coding, when it goes beyond building a landing web page or a simple mobile dashboard to calculate calories, can be very challenging and frustrating. Especially if, you like me, are not a professional coder.

The problem is mainly connected to the complexity and size of the app you want to build. The more features and functions you want to build into it, the more likely it is that:

• you will bump into memory issues,

• the AI will get stuck or

• get lost into interminable loops and that

• you will suddenly lose existing functionalities after you have worked hard at creating them.

So, as I tried to interpret these fallbacks as learning opportunities, I started strategizing more and more how to approach this type of coding and what steps to take to avoid as much as possible such negative situations.

The first lesson I brought home is that while it is absolutely essential to share with your coding AI a written brief of everything you want your little app to be, the best way to proceed is to develop one feature at a time.

At least, by proceeding in this way, I know what could be the cause of a new error and avoid falling into situations in which you have no idea of what caused what.

The second lesson learned has been that it is a good idea to first strategize with AI the path to be taken and only then to move to the execution part. So instead of just asking AI to do this or that, you first describe to AI what you have in mind to do, what the challenges may be and where you have eventually met problems before. By doing this you can then ask AI to suggest two or more alternative development paths, as well as the pros and cons of each.

This allows you to have a broader vision on how to proceed and to anticipate which parts of the work to be executed may bring unexpected issues.

The third lesson learned is not to trust your coding AI. This may not feel natural or intuitive in the beginning, but with time I realized it is the only way to come out alive from these projects.

In practical terms, not trusting your coding AI translates into verifying if you can ever step to a previous state of development / version of your app. Your AI will always tell you that YES, you can go back, but I learned the hard way not to believe everything AI told me. Can you really backup your work at discrete points and get back to them whenever you need to? What happens when you decide to do so? Better find out early in the process that when it could be too late.

This applies also to the many times where your AI coding companion will stumble on an apparently simple problem, and will repeatedly try to fix it but with no success. It is in these situations that you need to step in and question AI strategy, ask or suggest alternatives and force AI to think more before acting.

Finally, due to the many surprises, obstacles, losses of features you have worked hard at building, I learned that while being very exciting this work requires a lot of time, patience and the mind and attitude of a strategic warrior, or you will rapidly lose your temper and give up on what you had in mind to create.

When it comes to tools to do this type of work, I have been experimenting with several ones (Poe, ChatGPT, Claude, Gemini, SAME.dev, Github CoPilot and Bolt.new), but I have not yet found one that I would recommend as both usable and 100% reliable.

Bolt.new is the one that has behaved best of all the others tried so far, but even with Bolt I have fallen into infinite loops, into several problems that could not be solved after tens of tries, and by sudden losses of functionalities and features. Be prepared.

I will continue exploring other AI-coding platforms for my next mini-apps (Lovable, Replit, Cursor, V.0, Base44 and others) and will keep sharing what I discover along the way.

All of these AI coding tools are evolving by the day, and at the speed they are improving and getting better at this type of work, what it may not be done today or be an unsurmountable obstacle now, may become kids’ game by tomorrow.

Everything is in fast flux. And this is what makes this type of work so fascinating, intriguing and exciting, as you know you are working on the frontier of what is possible, and the frontier is moving with you.

The challenge has been much tougher than I expected, but the satisfaction is having now a working tool that I would have never dreamed of being able to create by myself.

The powerful realization is that we have definitely entered a new era, one in which, anyone who is serious and determined enough can develop and publish tools and apps according to his vision and ideas, without having to beg for VC funds or having to take on a full programming career.

Just like bloggers 20 years ago broke into the sacred publishing empire, giving to every individual the opportunity to become their own media company, today AI coding environments and LLMs open up the doors to everyone who wants to create value beyond words by creating new tools and instruments that are free from the VC-funded framework and from for profit-only corporate software development loops.

It looks like these are going to be great times for those wanting to change the world beyond writing about it.

More info on PersonalTab, where to download it and how to install it, here below, in the Premium section of this issue.