I think when you say “ditch localhost” you’re telling me to ditch my fast, instant-response laptop which I own and can peg the CPU of 24/7 for $0, in favour of a tiny cloud VM that I rent forever.

Your infra to run agents and builds for me is compared in my mind to a shell script an agent wrote a year ago and I reviewed once, that fires up my dev server and a local psql (5-10mb ram) on a dynamically allocated port hashed off the name of my current worktree, which it does so it doesn’t clash with other parallel work.

When the internet slows or dies I rarely notice.

As a cost conscious person who likes it when letters appear as I type, I think I might not be your ICP.

Am I being an asshole? Maybe. Am I going out of my way to tell you what goes on in the minds of people like me when we see offerings like this? Also maybe.

Also, do you see Boxes supporting OpenCode and self-hosted/local models in the future? If the rented machines have enough RAM and GPU access, it seems like there could be an interesting path toward a model-agnostic platform rather than being tied to the frontier labs.

I’ve been working on an [OSS TUI](https://github.com/prettysmartdev/awman) for managing agent execution and workflows in containers (local or remotely) and would love to collaborate if you’re interested.

Or, open source it and let us run it on our own VPS and keep your expensive cloud for those who want to pay. As it stands would never consider it.

If your CTO didn't spend the past year making an orchestration tool and a baby is he even qualified?

I have a vibe-coded orchestrator that I use to manage my claude and codex sessions across multiple machines, can also spin up sprites from fly.

https://github.com/tinkerer/propanes

warning: it is probably totally unsuitable for anyone else to use except for me

The main idea is a widget that you embed in your apps that lets you select elements, paste screenshots, and prompt what to change. This workflow is very productive for me. I would encourage everyone to add element selection to their orchestrators prompt composers. If you watch the looms on the readme note that my CLAUDE.MD calls me a Meat Computer and reminds me to hydrate.

I have a native tauri version that lets you select UI elements through the macos accessibility api too.

The session service uses tmux so you can open a native terminal via ssh and tmux attach. I add a ton of features that are in varying degrees of half-baked: the "brainstorm" mode allows you to do microphone transcription while interacting with the DOM and it will suggest tickets automatically. I've also been working on "bd2sdd" which is supposed to take your strings of user inputs and transform it into a spec, presumably because I also desired regressions. There are Wiggums (which aren't relevant anymore with /goal) and "FAFO swarms" (fan-out, aggregate, filter, optimze) which I use to reverse engineer other pieces of software, PowWow for codex and claude to work together.

I stole the structured views and remote session control from my friend's Agent Portal project txcl.io which is more fully-baked and narrower scope than propanes.

The ticketing system / tmux / structured views has been slowly evolving into multi-agent chat with a primary "Chief of Staff." It integrated pretty nicely into Slack.

- slow and outdated vms

- horrible/no way to standardize environments for my team

- no way to bring our own compute to help resolve these issues ^

I am building a self-hosted tool (OpenClaw-like) to solve the same problem (running agents 24/7 and access from monile), which I think is the main alterative approach to cloud tools. I'm glad that other people have recognized the problem.

We currently use worktrees btw. We have a port allocation system that sends ports to the agent automatically, which suffices for smoke testing web projects in parallel but requires some configuration. We've also found that asking agents to find a free port works as well. There's no way to get security-relevant isolation without a containerized system, but everything else can be worked around, and IMO more easily than the setup required to make a project ready for VM/container development.

Last time I tried to let AI build such extension, it told me that VSCode did not expose extension API to monitor AI chat.

edit: and if anyone knows of an existing service that has these properties, I'd love to know about it.

Why would I want this and not the other way around?

- I run hermes on the box and it has some scheduled cron jobs.

- I gave it an account on a custom Git forge. It cannot commit without my direct permission, though it can blow the setup up in other ways lol.

- I interact by assigning it issues and talking through Discord.