My Kindle was sitting unused, so I decided to see whether AI could assistance me discover a new use for it. I wanted to create a tradition app that could shield a range of uses and be controlled by an AI delegate on my MacBook.
I started alongside a small goal: use my own depiction as the Kindle's screensaver. Once that worked, I asked the delegate to build an app that could sequence through a gathering of photos.
We called the project KindleCTL. On my Kindle, I open "Photo Slideshow" from the identical display anywhere I choose my books. It cycles through ten photos, showing all for four minutes. When I desire to read, I exit the slideshow and come back to my books.
I used Codex and T3 Code on my Mac to activity alongside the coding agents.
I archetypal tried GPT-5.6 Sol alongside reasoning attempt set to Extra High. It wrote code to nexus to the Kindle but couldn't determine an error that prevented it from storing data on the device. Then I tried Fable 5.1 and Opus 5.5, but Anthropic's cybersecurity safeguards blocked all attempt following two prompts. Finally, I tried GPT-6 Astra alongside reasoning attempt set to High. We were capable to jailbreak the Kindle.
We used the existing jailbreak code as a citation to build our own tools for operating tradition application on the Kindle. Then we created and installed the slideshow app.
The slideshow passed its automated checks on my Mac, but switching photos on the Kindle remaining traces of the former depiction on screen. The tests hadn't caught how the bodily display refreshed. Getting from passing tests to item I could use turned out to be the most engaging part of the experiment.
Let's stroll through how we went from an unused Kindle to a operating photo frame, starting alongside getting the delegate connected.
Connecting the delegate to the Kindle
I started in an bare project directory alongside a 10th-generation Kindle Basic. I gave the delegate its network location and equipment details. We had no USB connection, so the complete setup happened complete Wi-Fi.
From there, we motionless needed to set up admission to run our own application on the Kindle.
The GPT-5.6 Sol attempt got as far as loading a helper on the Kindle, but the program that helper was meant to commencement never reported that it was prepared to execute. It couldn't find a filesystem anywhere it could shop content.
The attempts alongside Fable 5.1 and Opus 5.5 were the two blocked by Anthropic's cybersecurity safeguards following two prompts. Then I tried GPT-6 Astra. I explicitly asked it to build the tradition application we needed for individual use and obtain inspiration from existing open-source projects.
We adapted the existing jailbreak from SpiderCat, released by sparklerfish and the KindleModding community, and built tradition tools called KindleCTL about it.
- What we built: Tools to dispatch commands from my Mac, transfer files, and examine errors on the Kindle.
- What broke: Our launcher wrecked before it could start. We traced that to a compatibility matter alongside the Kindle's older scheme application and fixed how the launcher was built.
- How we checked: We connected alongside director access, transferred records the two ways, and displayed text. Later, we verified that admission motionless worked following a reboot.
The feedback iteration was the equivalent changer
Initially, the delegate could nexus to the Kindle but couldn't display data. To debug this, it asked me for pictures of the Kindle's screen. I took photos alongside my phone and shared them alongside the delegate on my Mac.
This was becoming tedious. Since the delegate was connected to the Kindle, I wondered why it couldn't get this data itself. I asked it to set up a feedback iteration so it could admission logs without me having to portion screenshots.
The delegate set up a persistent diagnostic association so it could collect errors and test fixes without asking me to download and reopen a bundle all time. Later, we added distant display captures so it could additionally inspect what the Kindle was displaying.
This was a turning item that let the delegate continue implementing changes on autopilot. Its reasoning traces showed it figuring out which folders survived a restart and which could shop external data.
I was amazed the delegate hadn't suggested this on its own. Once I asked for a feedback loop, it set one up and could test fixes without my help.
Most of the code was written in Python. The Kindle's launcher and slideshow were written in C, alongside casing scripts handling facility and equipment controls.
With the association working, I asked the delegate to try a small test first: transfer an depiction and set it as the Kindle's screensaver. I checked that the depiction appeared correctly and that waking the Kindle brought rear the normal Home screen. Once that worked, I asked for the slideshow app.
Getting the slideshow working
I asked the delegate for a looping slideshow of ten photos, alongside an Exit button. Each photo would fit the display during keeping its first aspect ratio.
The slides were changing, but the transitions weren't smooth, so I asked the delegate to appearance into it. With the feedback iteration already in place, it investigated the matter and used the Kindle's built-in display refresh scheme to apparent the former depiction following all glide change.
The display captures showed complete images, but they couldn't disclose traces remaining on the bodily display. After the refresh change, I checked the Kindle and confirmed that all photo looked clean.
I tailored how lengthy all photo stays on display and the brightness level. If the Kindle goes to sleep, the slideshow picks up anywhere it remaining off whenever it wakes.
For now, I use my Mac to alter the photos and slideshow timing. The photos remain stored on the Kindle, so the slideshow plant without Wi-Fi. In the future, we could add controls to alter the slideshow timing on the Kindle and dispatch photos immediately from a phone or another device.
What I learned
Next time, I'd set up the feedback iteration from the start so the delegate could peruse logs, inspect display captures, and test fixes on the device.
The Kindle now sits on our wooden shelf, anywhere we use it all day as a photo frame.
I started since I was inquisitive concerning what AI could assistance me do alongside an unused device. Now I have an app for my photos that plant the way I asked, affirmative the tools to keep experimenting alongside the Kindle.
Disclaimer: This was a individual test on my own Kindle, not a step-by-step jailbreak guide. Only test alongside devices you own or have definitive approval to modify.
