It’s 2026. People can use AI to generate copious amount of decent code, especially if they already program and especially if it’s for a low-stakes use case and imperfections are tolerable. And the world is awash in used tablets.
Thus it’s no surprise that at least 3 of my coder dad friends have mounted old tablets to the walls of their kitchens and living rooms and hacked together displays of information for their households. It replaces the paper calendars of yore with digital calendars, yes, but also displays family photos, weather forecasts, meal plans, shopping lists, chore lists, and various bespoke widgets.
A commercial version of such a product is the Skylight. But who wants to buy a new device and then pay a monthly fee? Everyone DIYs it instead. There must be thousands of takes on this project. But are they all one-off implementations of crufty code? And I can’t just deploy someone else’s solution? When I investigated the space six months ago for my own purposes, there wasn’t a single go-to candidate that looked usable. DAKboard is SaaS, Magic Mirror² runs on Raspberry Pi and a display?
EDIT —————-
After posting this, two other options were suggested to me:
Home Assistant: good fit if you’re in that world already! But too technically difficult for most people and not meant to run on an old tablet.
TRMNL: these devices are pretty cool and I like the mix of paid and self-hosted options. There’s not exactly a grid where you can plug anything in, like the Vercel-hosted dog coordinator app. Also not meant to run on an old tablet. But this was the closest thing to what I had envisioned so kudos to that project.
END EDIT —————-
It makes me think there’s a need for a standardized framework. I’m inexperienced when it comes to frontend development and have no mobile experience, so I’m not the ideal person to make it happen. But here’s my vision:
- A well-made, simple Android and iOS app that can run on the tablet 24/7
- It supports things like on/off/dim hours or waking-on-motion-detection
- It has a framework for building plug-ins. These are widgets that can display on a grid. Like dropping widgets on a smartphone home screen, they can be various sizes, moved around, and be assigned to new screens you can scroll to.
The plug-ins I would build would be for Google Calendar, Google Photos, a chore list for my kids, and a generic web view widget. My neighbor created an app deployed on Vercel where we coordinate which dog is outside so they don’t bark at each other through the fence.
(I think this latter part is where this product shines: you can plug in janky vibe-coded web apps via a single widget).
Other devs would contribute plug-ins for Home Assistant, Notion, shopping lists, weather, etc.
Apache Superset, the data visualization software I run at work and contribute to, has invested massive effort in refactoring the codebase to support extensions. VSCode is an inspiration in this regard. The idea is that extensions will make it way easier for developers to contribute new modular pieces to Superset. I’m thinking something similar here.
Then an end user doesn’t have to know coding to use it. They install from their app store, then select and configure the widgets they want. They just populate the necessary config and credentials for each widget.
Initially I wouldn’t worry about the “accessible to layperson end users” or “on the app store”, those have a higher bar for code quality and then someone is on the hook for maintenance (though maybe that can be monetized, especially if it provides API services like weather?). I would just make an app you can sideload (I have sideloaded onto Android before, apparently it’s possible on iOS too).
What I really want here is that instead of thousands of us doing a crappy job on this, we get one rock-solid framework app and then individual devs build high-quality plugins.
Does this already exist and I can use it & contribute instead of building from scratch? Please?
If not, who’s coming with me? What stack are we building this in? Where do we start?
