In recent weeks, developers and members of the user community have been employing AI agents to remove ads, preinstalled apps and tracking features from smart TVs, with the aim of creating simpler, faster, ad-free launchers and reducing data collection.
Notable projects and people
- The project dumbtv states on its GitHub: “Your TV should do what you paid for: your apps, your HDMI inputs, and nothing else. No ads from the TV itself, no watch tracking, no update you did not ask for.” dumbtv has been tested on several manufacturers’ TVs (Sony, TCL, Philips, Hisense, Xiaomi and Shield) and on Amazon Fire TV Edition devices.
- cleanyourtv was started by Gathin T.; he says the idea came one afternoon when he grew tired of the sponsored ads on his parents’ TV. In a tweet dated October 17, 2026, Gathin said Claude Code removed 25 Xiaomi packages over Wi‑Fi and installed a new launcher, making the set look “like an Apple TV.”
- Muzammil Yasin Sabuwala, the developer behind dumbtv, advises pointing an AI agent at the project’s GitHub repository to set it up. Sabuwala has described combining several AI models in his work and intends to extend model support over time. His tweet about the project appeared on October 18, 2026.
How the modifications work
These solutions often do not require root access to the TV and generally allow the user to restore factory settings if desired. Developers use AI models — including variants of Claude and other mentioned models — to identify installed packages, remove unwanted apps, replace the launcher and disable tracking features.
In practice, a user provides the TV model and network access to the AI agent, which then guides the modification process or produces commands. Multiple users reported removing dozens of packages (for example, 25) and replacing a stock 192 MB launcher with a custom 83 MB launcher, plus adding rotating wallpapers.
Why it matters
If such “dumb TV” modifications become widespread, they could affect the smart TV advertising ecosystem, which relies on selling viewing data and showing personalized ads on native interfaces. Disabling Automatic Content Recognition (ACR) and other tracking could reduce the data available to advertisers and change platforms’ revenue models.
Risks and limitations
These projects are works in progress and user reports are mixed:
- Some AI agents did not respond to simple prompts to “hack my TV,” requiring more elaborate instructions.
- Users warned that AI can make mistakes, potentially disabling important packages or breaking the system. One report said a user bricked their Android TV while using Claude; others became stuck partway through modifying a Fire TV.
For users reluctant to rely on AI, existing manual options for customizing Android-based smart TVs include ATV Launcher, Projectivity Launcher and FLauncher.
Conclusion
AI has lowered the barrier for a wider audience to alter the software on consumer electronics: many users are quickly removing ads and bloat from smart TVs. However, the tools carry technical risks, and broader adoption could have market, legal and security implications.
