On various Mozilla forums that I saw, the discussion was basically:
1. We can't just allow the browser to connect to any socket, since many either explicitly don't want browsers connecting to them, or are oblivious to browsers.
2. ...so we need to also add some sort of allow list
3. ...this is getting too complicated for such a niche feature.
So I think the nicheness was the high-order bit here.
i'm trying to understand how outer shell works here. on the website you give the following as your motivation:
> Apps like Jupyter and Tensorboard are not typically visible to standard web browsers if they’re running on remote servers, because it would be terribly unsafe to let the whole internet touch this app. Instead, they run on a local port on the server, which your computer can’t access directly.
> Classically, to get access to these, you had to open a new terminal and run:
is this true? isn't the normal thing just to do this ssh forwarding for prototyping, then for deployment, you set up a website like myjupyternotebook.com, and then set up auth so that others can't access it. HTTP basic auth is not too much work.
if you want SSH, not HTTP, to be what's publicly exposed, there's other options too, like putting it behind a VPN or tunnel.
all this to say, outer loop is super cool, but I don't get it. I must be missing something about why you built it, so could you help me understand?
I think there are different clusters of people who use servers, SSH, etc.
I'm closer to the cluster that uses them for deep learning experiments, GPU kernel optimization, robot development (a robot is just a server that moves!)... use cases where you are explicitly using a remote computer.
For this cluster of people, I think this tool feels more intuitive than the flow you suggest. But maybe I'm projecting!
And, to me, this just feels like one of the fundamental things that could exist; it's like a graphical operating system, but remote-first.
I guess it saves you the hassle of dealing with reverse proxies and TLS certs if your use case is "userbase is 1 person and it is me, and i only access services from a desktop os"
Lovely writeup! I'll bookmark this for my own research.
My terminal's "clickity clackity" features [0] are local to the machine so I lose graphical-ness as soon as we remote in somewhere.
That's starting to change a bit with offline replay [1] where the native GUI and TUI work in tandem to unlock some rewind. But there's quite a road ahead and I love seeing others experiment properly. (Terminals are massively underserved.)
Being able to initiate a shell app from a regular remote ssh CLI prompt (like "ApacheConfig myhost.com" or "Editor ~/myrepo") might improve integration with people's existing CLI workflows.
It does need an agent that starts with every X or Wayland session and waits for requests from remote SSH sessions to start an app.
I’m good with just tailscale and self-hosted web-apps. Seems the main selling point is either native UX or reduced barriers to entry security-wise. I like barriers to entry.
I don't really know what outerframe frame is. I tried to understand from the video and the blog but I'm still not sure what it is. Is it like a web browser but instead of DOM, HTML and JS you have Swift and SwiftUI running in a sandbox?
If so how would that work on non Apple devices? Also how much will that sandbox protect you?
It's a fun heretical idea, moving away from a "cross-platform" web to a "multi-platform" web. It's a cross-platform protocol that hands off to platform-specific frontend code. I think it's a natural direction for the web, in a world where LLMs can translate to other platforms.
That's interesting idea. If we put into CLI with some ANSI escape code, that may become something real. Imagine a normal terminal app just render part of the UI in web and communicating in UNIX socket. While doing the fancy UI, everything is still controllable with keyboard, and optionally with mouse. The UI will fallback to text UI for older terminal
No no not something on top of the UI stack. They also need framebuffer support so they are big headache to setup on headless server.
What I mean is that we can bring some web tech to terminal natively. We don't even need a separated shell. Security and bi-directional communication is built by default because of UNIX socket. But we still need to think how to handle stuff like cookie, local storage, external CSS / JS, ...
Just had a quick look but I like the look so far. I’ve been thinking along similar lines for ages but never quite got around to making something. I very much support any effort to make remoting less dependent on the archaic character grid.
I wrote an early version of the Cylance AV desktop client. The UI side was a web app that talked to its windows service backend using HTTP over windows pipes. This was surprisingly easy to do using WCF.
One of the more interesting pieces of Microsoft software is the Windows Admin Center where it's a web app to configure a Windows Server. Ideally, it was made for core installs where there's no GUI but it's there as a viable web management panel.
The tool from OP and WAC are pretty similar in terms of functionality and usecase. Why would you want this? Well, imagine your team needing to be able to do server functions but you have less technical team members to do it for you, which is very often the case in big places, most people are familiar with the web browser and having a website to do these sorts of actions makes it easier to have things done in one place without a lot of tools like Remote Desktop, SSH, WinRM, etc. configured.
As someone managing various servers, both at home and at work, I see how this can be really useful. I see it not in the production space yet but rather in the experimenting, using a Linux machine as a second compute device!
So regarding your last point, I'm convinced. I think it is useful!
The one fact that is bugging me is that now it requires a client specific app, with GUI, on my PC and I wonder if using ssh port forwarding could reduce the surface. I mean I wonder if either having a rich client that executes commands via ssh or a rich server (including Web Server) with ssh port wouldn't suffice, so that I can avoid installing stuff on the server AND on my computer.
I thought this looks interesting, but was a little confused with what appears to be MacOS-only support at https://outerloop.sh/? I'm running Ubuntu 24.04, I kind of assumed from context that it'd be something I could spin up in a few minutes just to give it a go?
Also worth noting, my decision to give it a go relied mostly on the fact that I couldn't quite work out what the product is. Having "Outer Shell" and "Outer Loop" described as distinct-but-connected entities is a little confusing, IMO, which do I need to install, on what, and in what order?
I can’t make up my mind if I love it or hate it. On one hand this is like SSHapi on the other there’s no structure, no contract… i had similar doubts with Cockpit.
I'm confused -- does this compile it live when the server ships code?
How do we resolve dependencies, toolset etc..
Is the idea to just pick an old enough platform toolchain you expect to be present?
In all cases, the code is pre-compiled. A user never waits for anything to compile. When Outer Loop installs Outer Shell, it downloads pre-compiled binaries to the server. For Linux these are compiled against a manylinux ABI. Ditto for when Outer Shell installs one of the bundled apps. When a backend serves a native "web" app over HTTP it sends already-compiled ARM (or x86) code to the client.
Dependencies are less of a concern for the frontend binaries. For backends, I use a dependency-light approach, static-linking anything that's needed. Of course, people are welcome to do backends however they want, and just tell Outer Shell about the systemd/launchd units via the API. I used this no-dependency approach to keep everything lightweight and to keep install steps trivial, but admittedly it pushes me in certain directions (for example, using custom binary formats rather than sqlite).
I am not sure I'd use this over exposing websites with wireguard as those will automatically work across platforms. But it looks like you could create some really cool experiences with it, and I'm happy people are exploring this space.
I don't have a dog in this fight, and anyway dogfighting is bad, but the intro to the Wikipedia article[0] reads:
> An operating system shell is a computer program that provides relatively broad and direct access to the system on which it runs. The term shell refers to how it is a relatively thin layer around an operating system.
> Most shells are command-line interface (CLI) programs. Some graphical user interfaces (GUI) also include shells.
The last line I think supports the notion that the term "shell" at least implies a CLI, but I can understand both positions.
The earliest versions of MacOS, all the way up through 9, had a ROM call at 0xA9F4 which was labeled `_exitToShell`. In the days before pre-emptive multitasking, this instruction's job was to force the current application to close and return the user to the MacOS desktop (the Finder). The "shell" in this context being the desktop user interface.
I wondered if this would be controversial. It all depends where you grew up.
> Cairo, like Chicago, had a new shell (Microsoft’s favorite word for the user interface for launching programs and managing files) and a new file system
Not really no. I’ve been using shells and authoring new ones for around 40 years across a variety of platforms. The term has always been pretty loosely defined because as technology evolved the term “shell” was borrowed. So like I said, a shell can refer to a graphical core just as much as a text-based one. You can get webshells too.
The original intent was that a shell is a thin wrapper on top of the OS to expose the hosts capabilities. But that hasn’t been an apt description for most of those 40 years.
command line shell vs graphical shell. My first experience with a graphical shell was dosshell[1]. For a while we called the Windows 3.1 interface "the shell". I guess the terminology has changed since that time.
This appears to me like a solution in search of a problem, like many others before it...the quote below seems relevant to this effort.
"Those who do not understand Unix are condemned to reinvent it, poorly." ~Henry Spencer
I like the idea of separating the frontend and backend of a graphical app. But I feel like this is hardly a novel idea, maybe I'm missing something.
I take it you don't know about "X11Forwarding yes" or "html5 web app"
That is a security concern, that's why it isn't implemented. At least raw unix socks. You can have WebSockets and other ports only limited to http.Quick response regarding security:
On various Mozilla forums that I saw, the discussion was basically: 1. We can't just allow the browser to connect to any socket, since many either explicitly don't want browsers connecting to them, or are oblivious to browsers. 2. ...so we need to also add some sort of allow list 3. ...this is getting too complicated for such a niche feature.
So I think the nicheness was the high-order bit here.
(FYI, Outer Loop does add an allow-list: https://outerloop.sh/unix-domain-sockets/)
i'm trying to understand how outer shell works here. on the website you give the following as your motivation:
> Apps like Jupyter and Tensorboard are not typically visible to standard web browsers if they’re running on remote servers, because it would be terribly unsafe to let the whole internet touch this app. Instead, they run on a local port on the server, which your computer can’t access directly.
> Classically, to get access to these, you had to open a new terminal and run:
> ssh -L 24601:localhost:8889 mrcslws@lambda4.mycompany.com &
> ssh -L 24602:localhost:6006 mrcslws@lambda4.mycompany.com &
is this true? isn't the normal thing just to do this ssh forwarding for prototyping, then for deployment, you set up a website like myjupyternotebook.com, and then set up auth so that others can't access it. HTTP basic auth is not too much work.
if you want SSH, not HTTP, to be what's publicly exposed, there's other options too, like putting it behind a VPN or tunnel.
all this to say, outer loop is super cool, but I don't get it. I must be missing something about why you built it, so could you help me understand?
I think there are different clusters of people who use servers, SSH, etc.
I'm closer to the cluster that uses them for deep learning experiments, GPU kernel optimization, robot development (a robot is just a server that moves!)... use cases where you are explicitly using a remote computer.
For this cluster of people, I think this tool feels more intuitive than the flow you suggest. But maybe I'm projecting!
And, to me, this just feels like one of the fundamental things that could exist; it's like a graphical operating system, but remote-first.
I guess it saves you the hassle of dealing with reverse proxies and TLS certs if your use case is "userbase is 1 person and it is me, and i only access services from a desktop os"
Ever since I started using Caddy, doing that has been soooo easy.
Download the binary, make a Caddyfile
And then just "./caddy start"does this work with multiple caddy servers? ie can you bind multiple caddy servers to port 80/443?
You can have multiple configs in a single Caddyfile and reload when you make changes, and it'll just route them as you wish, e.g.
domain1.com -> service on port 1234
domain2.com -> service on port 5678
domain3.com -> serving a file directory.
And then you still access domain1.com, domain2.com, domain3.com on port 80/443
You set up multiple services behind a single caddy reverse proxy
Lovely writeup! I'll bookmark this for my own research.
My terminal's "clickity clackity" features [0] are local to the machine so I lose graphical-ness as soon as we remote in somewhere.
That's starting to change a bit with offline replay [1] where the native GUI and TUI work in tandem to unlock some rewind. But there's quite a road ahead and I love seeing others experiment properly. (Terminals are massively underserved.)
[0] https://terminal.click
[1] https://terminal.click/posts/2026/06/tui-stability/#:~:text=...
Being able to initiate a shell app from a regular remote ssh CLI prompt (like "ApacheConfig myhost.com" or "Editor ~/myrepo") might improve integration with people's existing CLI workflows.
It does need an agent that starts with every X or Wayland session and waits for requests from remote SSH sessions to start an app.
I’m good with just tailscale and self-hosted web-apps. Seems the main selling point is either native UX or reduced barriers to entry security-wise. I like barriers to entry.
I don't really know what outerframe frame is. I tried to understand from the video and the blog but I'm still not sure what it is. Is it like a web browser but instead of DOM, HTML and JS you have Swift and SwiftUI running in a sandbox?
If so how would that work on non Apple devices? Also how much will that sandbox protect you?
It's purportedly cross-platform. The documentation leaves a lot to be desired, but it is described more here:
https://outerframe.org/ and https://outerloop.sh/native-apps/
Also a blog post about it, with its own video: https://probablymarcus.com/blocks/2026/05/10/like-a-web-view...
It's a fun heretical idea, moving away from a "cross-platform" web to a "multi-platform" web. It's a cross-platform protocol that hands off to platform-specific frontend code. I think it's a natural direction for the web, in a world where LLMs can translate to other platforms.
That's interesting idea. If we put into CLI with some ANSI escape code, that may become something real. Imagine a normal terminal app just render part of the UI in web and communicating in UNIX socket. While doing the fancy UI, everything is still controllable with keyboard, and optionally with mouse. The UI will fallback to text UI for older terminal
So, uh... X11? VNC? RDP?
No no not something on top of the UI stack. They also need framebuffer support so they are big headache to setup on headless server.
What I mean is that we can bring some web tech to terminal natively. We don't even need a separated shell. Security and bi-directional communication is built by default because of UNIX socket. But we still need to think how to handle stuff like cookie, local storage, external CSS / JS, ...
Just had a quick look but I like the look so far. I’ve been thinking along similar lines for ages but never quite got around to making something. I very much support any effort to make remoting less dependent on the archaic character grid.
I wrote an early version of the Cylance AV desktop client. The UI side was a web app that talked to its windows service backend using HTTP over windows pipes. This was surprisingly easy to do using WCF.
One of the more interesting pieces of Microsoft software is the Windows Admin Center where it's a web app to configure a Windows Server. Ideally, it was made for core installs where there's no GUI but it's there as a viable web management panel.
The tool from OP and WAC are pretty similar in terms of functionality and usecase. Why would you want this? Well, imagine your team needing to be able to do server functions but you have less technical team members to do it for you, which is very often the case in big places, most people are familiar with the web browser and having a website to do these sorts of actions makes it easier to have things done in one place without a lot of tools like Remote Desktop, SSH, WinRM, etc. configured.
Lovely video and ingenious implementation. Kudos!
As someone managing various servers, both at home and at work, I see how this can be really useful. I see it not in the production space yet but rather in the experimenting, using a Linux machine as a second compute device!
So regarding your last point, I'm convinced. I think it is useful! The one fact that is bugging me is that now it requires a client specific app, with GUI, on my PC and I wonder if using ssh port forwarding could reduce the surface. I mean I wonder if either having a rich client that executes commands via ssh or a rich server (including Web Server) with ssh port wouldn't suffice, so that I can avoid installing stuff on the server AND on my computer.
I thought this looks interesting, but was a little confused with what appears to be MacOS-only support at https://outerloop.sh/? I'm running Ubuntu 24.04, I kind of assumed from context that it'd be something I could spin up in a few minutes just to give it a go?
Also worth noting, my decision to give it a go relied mostly on the fact that I couldn't quite work out what the product is. Having "Outer Shell" and "Outer Loop" described as distinct-but-connected entities is a little confusing, IMO, which do I need to install, on what, and in what order?
Cool idea anyway, no shade here.
I can’t make up my mind if I love it or hate it. On one hand this is like SSHapi on the other there’s no structure, no contract… i had similar doubts with Cockpit.
This is cool. Though I don't see why someone would want to do more work/design for the custom GUI rendering for a custom/renderer (your viewer app) ?
I'm confused -- does this compile it live when the server ships code? How do we resolve dependencies, toolset etc.. Is the idea to just pick an old enough platform toolchain you expect to be present?
In all cases, the code is pre-compiled. A user never waits for anything to compile. When Outer Loop installs Outer Shell, it downloads pre-compiled binaries to the server. For Linux these are compiled against a manylinux ABI. Ditto for when Outer Shell installs one of the bundled apps. When a backend serves a native "web" app over HTTP it sends already-compiled ARM (or x86) code to the client.
Dependencies are less of a concern for the frontend binaries. For backends, I use a dependency-light approach, static-linking anything that's needed. Of course, people are welcome to do backends however they want, and just tell Outer Shell about the systemd/launchd units via the API. I used this no-dependency approach to keep everything lightweight and to keep install steps trivial, but admittedly it pushes me in certain directions (for example, using custom binary formats rather than sqlite).
Interesting, kind of like a more fancy web shell. Haven't really ever seen the need for those, mostly because terminals work better than browsers.
Sometimes the browser is the only "computing platform" you have available (e.g. on some mobile devices, hotel kiosks).
I am not sure I'd use this over exposing websites with wireguard as those will automatically work across platforms. But it looks like you could create some really cool experiences with it, and I'm happy people are exploring this space.
I prefer hytelnet and MUDs but I don’t count, I’m just too old.
> every app is a small HTTP server
This adds unnecessary overhead for communication. using web and web-like approaches on desktop system is a terrible idea.
Defeats the purpose of the shell. The shell is for CLI interaction.
No. A shell is any user interface. Windows shell is explorer.exe and it used to be possible to change that via a config line in a system INI file.
SSH protocol also isn’t just for CLI work. It supports file transport (eg SFTP), TCP/IP forwarding and even SOCKS HTTP proxying.
You also used to be able to run GUI applications over SSH via X11.
You have a very loose definition of a shell that conflicts with about 40 years of history.
I don't have a dog in this fight, and anyway dogfighting is bad, but the intro to the Wikipedia article[0] reads:
> An operating system shell is a computer program that provides relatively broad and direct access to the system on which it runs. The term shell refers to how it is a relatively thin layer around an operating system.
> Most shells are command-line interface (CLI) programs. Some graphical user interfaces (GUI) also include shells.
The last line I think supports the notion that the term "shell" at least implies a CLI, but I can understand both positions.
---
0. https://en.wikipedia.org/wiki/Shell_(computing)
Edit: I'm shite at formatting on HN
The earliest versions of MacOS, all the way up through 9, had a ROM call at 0xA9F4 which was labeled `_exitToShell`. In the days before pre-emptive multitasking, this instruction's job was to force the current application to close and return the user to the MacOS desktop (the Finder). The "shell" in this context being the desktop user interface.
Just FYI.
I wondered if this would be controversial. It all depends where you grew up.
> Cairo, like Chicago, had a new shell (Microsoft’s favorite word for the user interface for launching programs and managing files) and a new file system
https://hardcoresoftware.learningbyshipping.com/p/020-innova...
When I worked at Microsoft 2010 - 2014, the word "shell" was still used in this way. I decided to say "graphical shell", to make it clearer.
Not really no. I’ve been using shells and authoring new ones for around 40 years across a variety of platforms. The term has always been pretty loosely defined because as technology evolved the term “shell” was borrowed. So like I said, a shell can refer to a graphical core just as much as a text-based one. You can get webshells too.
The original intent was that a shell is a thin wrapper on top of the OS to expose the hosts capabilities. But that hasn’t been an apt description for most of those 40 years.
Appeal to authority.
command line shell vs graphical shell. My first experience with a graphical shell was dosshell[1]. For a while we called the Windows 3.1 interface "the shell". I guess the terminology has changed since that time.
[1] https://en.wikipedia.org/wiki/DOS_Shell
> Isn’t it weird that this doesn’t already exist?
It does. MobaXterm have a bunch of it already, file manager on the side and ability to pass X11
Edit: withdrawing this objection, had no idea that right-clicking allowed the speed to be adjusted.
Sure, I just added YouTube mirror link to the post: https://youtu.be/e40PLLuZ5KI
(The one on the website is the standard browser video player, not custom.)
Thanks (and to pelzatessa as well), TIL about the right-click menu on these. That'll come in handy.
but its just standard <video> element, in firefox I can even right-click to change the speed to 2x. It's certainly better privacy-wise.
Heresy!