Pane – An open-source browser that turns your work into living websites

TLDR: I wasn't satisfied with any of the AI browsers available right now. So, I went wild and built exactly what I wanted. I built a browser that's as comprehensive as an autonomous agent like Hermes, while removing the need for hundreds of integrations since this is literally your browser.

Then I thought, what if the browser could build websites that are built just for my work, and can grow with my work. So my workflow changes from chatting with a bot to navigating these websites and clicking buttons for actions built just for me.

A simple eg: a sales pipeline tracker, that keeps you prepared for the next meeting, and automatically stores the details of the previous interactions and details for all the prospects.

So, here's Pane.

Features:

- Usual AI browser with AI and automation

- Powers of an autonomous agent like Hermes, or OpenClaw -> soul, memory, self-improving skills etc.

- This also solves a major problem with Hermes -> integrations. I had to have multiple integrations to make Hermes useful for anything. Since Pane is literally my browser, search, access, auth etc are already solved. So I almost never need an integration.

- Built in meeting transcription.

- Auto context capture and learning from your browsing patterns. So you can say things like, "resume my research work from yesterday".

- The flagship: Personalised internet. Chats are poor containers for work that lasts days or weeks. Pane creates private, living sites for it instead: a Job Search pipeline, a Research Hub, a Sales Pipeline, or a site shaped around what you are doing.

Story: I have been trying AI browsers since Comet was invite-only. Since then I've tried Atlas, Browseros, Dia, and most recently Aside (came closest to what I wanted). But none have came close to give a real productivity boost.

I've also been an avid Hermes user, and it has literally been Chief of Staff for all my startups. I've been able to automate 30% of my GTM work with hermes, but it always fell short because I'll have to keep hacking connections to my browser because it has to use my accounts for integrations such as email outreach analysis, linkedin outreach, etc. This is partially the reason I've been trying AI browsers.

So a month ago, I gave in, forked the best bedrock that was available, BrowserOS (it's a good browser btw, kudos to the creators for the good work). Went wild to add everything I wanted out of my browser.

Then, I got the idea of Personalised internet: since the browser already knows all of the work, it can keep building personalised websites for my work. For example: a dashboard for GTM work progress, a kanban board for my social media outreach with details of each outreach and so much more. This way, I don't have to keep asking the agent, "where are we with X", I just open the personalised websites and navigate around. The backend is the agent, so buttons are prompts -> the agent decides what happens next when you click something, like load another dynamic page, do an action etc.

To be honest, I've been loving it. Releasing it to the internet to see what the rest of the world thinks of it.

It's still very early, and there's so much more possible with this, and I'll keep making this the OS for my work. Would love to have your feedback if you find it interesting and end up using it.