Hmm, I haven't thought of this use-case. I think this could also work if you put the registry entries into HKEY_CURRENT_USER (as opposed to HKEY_LOCAL_MACHINE).
If you'd like to experiment, you can change the register script to put entries into HKCU instead of HKLM. It might just work!
The uninstall form uses Google Analytics to record answers. I suspect you've configured your browser to block tracking content, which is why it doesn't work for you. Thanks for letting me know this - I've updated the error message so that other people would understand why it doesn't work.
As for your actual feedback: there is no way to achieve this from the extension. The extension doesn't know where the companion executable is - it is the browser that connects the two applications.
However, you can enable debugging, then you can open the companion log file, which will contain the path to the executable.
Out of curiosity, why do you need to know the location? What were you trying to change in the registry? Is this the reason you uninstalled it? Maybe I can help.
I'm assuming by "using the window manager in Linux" you mean you can manually put windows always-on-top.
Three reasons this is different:
1. The window is actually not on top of every window - it's only top of your browser. It's "attached" to your browser window, same way a real PiP would.
2. It's less clicks and also positions the window automatically for you.
3. It's cross platform. Windows doesn't have an easy user-accessible way to enable always-on-top.
I knew these tiling managers exist for tabs, but I wanted to take a slightly different approach on multitasking. TabFloater is primarily useful if you use your browser in a single-window mode. Within one window, it's impossible (AFAIK) to arrange tabs the way you mentioned, you would need separate windows. If you'd like that functionality, the above linked extension should work great.
I'm aware that Windows User Access Control blocks the Companion installer - sadly I don't have a way to fix this at the moment. I would need to sign the installer exe with a code signing certificate that is trusted by Windows, and this costs hundreds of dollars. I can't afford this.
As for the usage stats: if enabled, it sends data to Google Analytics, stuff like how many people installed it, how many people used X or Y feature. It really only exist to help me track how people use it.
It is of course completely up to you if you want to disable it - but as with any product, usage stats just help developers make the product better.
No, because as soon as you click on "large" browser window, the smaller one will drop to the background immediately. It won't stay on top.
You could of course move the window next to your main browser window - but, you would need to first open a new window, move your tab into that window, resize and position it manually... it's tedious.
I have the same resolution size :) I agree that the usefulness of this app depends on how you use your browser. I do the exact opposite of you, at any time I have at least 20-30 tabs open and I regularly switch back and forth.
I prefer to think of TabFloater as not an app that creates/moves/manipulates extra browser windows (even though that is what actually happens), but an app that adds an extra browser functionality (picture-in-picture). In this sense, I think it makes sense that it brings its own "always-on-topness".
I'm also aware that this app solves a pretty niche use-case - but it was fun developing it and I learned a lot!
Depends on what you mean. If we are talking about whether or not TabFloater itself can be trusted, then I'm 100% confident since I wrote it :) It's fully open source, the companion binaries are built through GitHub Actions, so I have no chance of introducing malicious code. The Chrome extension is also published through GitHub Actions. All this means that you can be sure you're running the same code that you read on GitHub. Publishing the Firefox addon is manual at the moment (they don't offer a great API for automated publishing), but Firefox requires me to submit source code alongside the extension, and they review it.
If you meant an unintentional security issue, then I think we are still good. The only thing the extension communicates to the native application is the title of the current browser tab, which is arguably personal information. However the native application is running on your machine, and you chose to install it. There's no way for remote machines to get your info this way. The only scenario I can see for leaking private info is if someone creates a malware that mimics TabFloater Companion and somehow manages to get it on your machine without you knowing. Then, they will get the titles of the sites you use TabFloater on. There are so many better ways for attackers to get the same information - I think this risk is negligible.
Of course, if you see any other issue that I might have missed, do let me know!
Indeed, I did not found this particular thread, but I have come across the fact that it was supported in the past. I guess my solution will also be a bit fragile, but so far, it works :)
I was also contemplating removing the borders and title bar of the floating window, so that it would much more like a panel, but on Windows, Chrome actually draws its own title bar, so you cannot remove it using the Windows API.
Hello everybody! I created a browser extension called TabFloater, which can replicate the picture-in-picture functionality with any website. It moves a chosen tab into a smaller window, which is going to remain always on top, so you can interact with both websites at the same time.
This is my very first post on HN. I have been working on this pet project for a bit over a year now. The motivation for it was simple: because I use my browser in a single, fully maximized window, many times I found myself switching back and forth between two tabs, for example to copy something from one website and use it on another. I know there are many extensions that can rearrange tabs for you, but I wanted something a little bit different. I really liked how picture-in-picture works for videos - I wanted to do the same!
Easy enough, I thought. I just need to create a new browser window, move the given tab into it, then... I just need to make it always on top. Right? Well, turns out, the browser API doesn't allow you to do that. So after a lot of trial end error, I came up with a way to achieve it: I implemented a small app in C++ that can manipulate the browser window and set the always-on-top property. Then "all I needed" was to hook it up with the browser extension and voila.
I also implemented a feature I call "Smart positioning": if you enable it, the extension will analyze the DOM structure of a given site and tries to identify the largest empty area and puts the floating tab there. This way, it doesn't cover up any text, pictures or videos. It more or less works :)
Sounds pretty simple, but I spent a LOT of time trying out different things. I use Linux, so naturally I implemented it for that platform first, but I knew it would never be useful for other people if there wasn't a Windows version as well - so I did that too. I worked on this project on evenings and weekends, and I learned a lot. This was my first proper pet project and I'm really glad I was able to see it through!
The extension is available for Chrome and Firefox, and you can use it on Windows and most Linux distros. The native application doesn't have an OS/X version - sorry :) I don't own a Mac and I don't know anything about the Apple ecosystem. I'm looking for help in that department, if you're interested! :)
Reminder: in addition to the browser extension, you must also install the native application (called Companion).
If you'd like to experiment, you can change the register script to put entries into HKCU instead of HKLM. It might just work!