Read-it-later apps only helped organise the clutter, creating neat filing cabinets for things I still did not read. The core problem remained: how to turn raw information into actual understanding, and to do it in my own tailored way.
So I built my own tool using AI (obviously). It is a simple web app that serves as my personal reading space. It pulls bookmarks from Readeck to start, but inside, the entire experience is mine. The interface is built for focus, with customisable fonts and themes that make it comfortable enough for long reads on my e-ink tablet.
While it remains a place for reading first and foremost, I have integrated simple tools using local Ollama models. They are quiet helpers, available only when I call on them, allowing me to select a dense paragraph for an explanation or generate a quick summary of the whole piece.
I recently added local text-to-speech, so I can now listen to an article or just its summary. This is a useful option for moments when I’m moving around, like during cooking. It’s similar to ‘freemium’ apps such as Elevenlabs. Only here I’m in control of everything. The app itself and all models run on my desktop, which makes it fast and completely private, yet I can still access it from anywhere—my phone, my tablet or my laptop.
Building it has been as rewarding as using it. There is a real satisfaction in co-crafting a tool that fits the unique contours of your own mind, rather than forcing your workflow to fit a pre-packaged product. It is my answer to the noise, and my way through the pile.