Founder note

Why I built sideBar

Most AI products still make people move context between tools, copy and paste into chat, and rebuild the same setup every time they want help. That copy-paste loop is the wrong shape for something that is supposed to save time.

I kept running into the same friction: the model answer might be good, but I had to do too much manual setup to get there. The work lived in one place, the assistant lived in another, and the distance between them became the product.

Claude Code made the idea click for me. It starts in the right environment, already surrounded by the files and context it needs. That changes the experience completely, because the assistant is not starting from a blank box. Personal AI should work more like that.

sideBar started as my answer to that problem: a place where notes, tasks, saved pages, and the assistant stay close together so help begins from the context that is already there.

I became more and more convinced that most AI products still have the shape wrong. They are existing apps with AI bolted onto the side, when the more interesting opportunity is to build around an AI foundation from the start.

I built sideBar for myself first. It started as a read-it-later tool because that was the job I needed most. Then I added notes. Then tasks. Over time, the things I used to spread across different apps started to come together in one place.

That is the idea behind what I think of as AI-native productivity. The assistant should not be a detached chat box you visit separately. It should live inside the work surface itself, with access to the right context at the moment you need help.

That also shaped the kind of assistant I wanted to build. I did not want a generic AI shell. I wanted sideBar to have a personality, a point of view, and enough self-awareness to help people understand what it can actually do instead of dropping them into a blank box and leaving them to figure it out.

Skills are a big part of that. They are how sideBar grows over time. Some capabilities are built in, others can be added when they are genuinely useful, so the product can expand without turning into a messy collection of disconnected AI tricks.