
His parents wanted to open Netflix. Their TV wanted to show them an advert first.
Developer Gathin says he used Claude Code to sort it out. He shared the work as an open-source project, originally called tv-vibehack and now CleanYourTV.
This is the kind of AI story I like. A familiar annoyance, something you can actually change, and code other people can inspect.
I am Alex. Welcome to ShortCu8 by Innov8.
Let's dive in.
What did AI actually fix?
Claude Code connected to his parents' Xiaomi TV over their home network, disabled unwanted software and built a simpler home screen.
The TV's display hardware stayed the same. The change was to the software getting between his parents and what they wanted to watch.
A “launcher” is just the home screen where you choose your apps. Replacing it can give you a quieter way to reach them. Cleaning up unwanted apps is a separate job.
CleanYourTV includes both, along with model-specific examples and restoration tools. Its main instructions target Android TV and Google TV, with limitations on other systems. This does not remove adverts inside streaming services. Project documentation
So before anyone starts celebrating free YouTube Premium: wrong kind of ad removal. 😅
The tools existed. Now you can ask for the work.
These projects use Android's developer connection, called ADB. After you enable it and approve the computer on the TV, software on that computer can inspect and change parts of the TV's setup.
The AI is helping operate those tools. You need an assistant such as Claude Code or Codex running on a computer with the right access, rather than an ordinary chat window that only gives you advice. TV Detox's explanation
That is why the example is interesting beyond televisions. You can describe the problem in ordinary language, while the assistant helps translate it into technical steps.
Of course, being able to run a command is different from knowing whether it was a good idea.
Three places to start
1. CleanYourTV: the project behind the story.
Browse the repo. It contains the launcher, setup prompt and contributed TV examples. Read the notes for your exact model and firmware rather than assuming a similar brand name is enough.
2. Mert Cobanov's guide: an explanation before the commands.
Read the guide or browse its GitHub repo. It walks through connecting a TV to Claude Code or Codex. Its prompt asks the assistant to save the existing setup, measure memory use and keep a record of changes. Useful if you want to understand the process before committing to it.
3. TV Detox: another cleanup prompt and small home screen.
Browse the repo. It includes an undo script and a linked demonstration on a real TV. You can also keep your current home screen and use its cleanup approach separately.
I have checked these projects' documentation, not tested them on your television. Treat the instructions as something to review for your device.
Keep a way back
Cobanov describes accidentally disabling a package that controlled his TV's Inputs button. The name did not make its purpose obvious. His guide also says to get the replacement home screen working before disabling the original. Guide and recovery notes
That is a more useful lesson than “paste this prompt and relax.”
Start with the smallest change. Have the assistant explain what it plans to disable, record how to undo it, and test the remote, sound, HDMI inputs and your streaming apps afterwards. Turn debugging off when the work is finished. Setup and verification approach
There is something satisfying about this example. Nobody in that living room needed a more impressive AI demo. They wanted the TV to open the app they selected.
For once, the interesting technology was the part helping the other technology get out of the way.
The ShortList
🛠️Cool Tools of the Week:
GPT-Rosalind: OpenAI launches new Rosalind Academic Plan
Flux 3 Image Klein: Black Forest Labs' open weights with quality of a large frontier model
Chrome tabs: ChatGPT can now use your tabs in Chrome to make answers more useful
📩 Innathe Shortcu8 engane undarunnu 👇️?
We read every reply - just reply to this email and let us know how we can improve !
Appo adutha Shortcu8il kanaam bie…👋
If you read till here, you might find this interesting
#AD1
State of Product reveals what AI still hasn’t solved
80% of product professionals say AI helps them ship faster, but customers aren’t seeing value any sooner. Atlassian’s State of Product 2027 explores the gains AI is delivering and the challenges that remain, from decision-making that hasn’t kept pace to gut instinct overriding customer evidence.
#AD2
Turn today's calls into tomorrow's proposal
Wispr Flow Notetaker brings your meetings into Claude or ChatGPT through a built-in connector. Ask for a proposal draft or a deal recap, and your AI works from an accurate record of what was said.







