Show HN: Ultrafast Rust Based Agent Workspace (trypalomino.com)

10 points by gpt-7 •

Excited to show Palomino, our GPUI based ADE!

Palomino is very fast and lightweight

A lot more is coming soon and we appreciate all feedback!

Does it use libghostty-vt?

"Palomino does not send your code, files, prompts, or chats to us. AI services you choose to use receive the content you send them."

Wouldn't "AIDE" be more skimmable? Took me a while to figure out you probably mean something like "agentic development environment" because I couldn't find a definition of the acronym anywhere.

how does it compare to https://herdr.dev/ ?

A lot of our users love the UI, which we’ve spent a lot of time working on!

In addition to terminals, Palomino lets you create tiles for browsers, simulators, and agent chats.

Also coming soon. We also are looking at cloud machines and remote machine management in the app.

Tangential question: is getting try/use<something>.com gets an auto boost from HN?

Ref: https://news.social-protocols.org/stats?id=50310055

Or is it a Rust boost?

These days you can just make your own? I can strongly recommend everyone just build their own personal tool...

Or am I missing something?

Here is my personal, Rust based, ultra-fast ADE. Feel to fork an make it your own:

github.com/rjprins/agtk

Tmux, nvim, opencode + a load of custom scripts and skills is all I need. I cringe every time I see some colleague daddle around with sth like this or copilot trying to phrase a prompt what should be just a git command. But I might just not be the target audience.

Exactly. Mine is Ghostty + Kakoune + Nushell sprinkled with Apple Script

Ref: https://github.com/paaloeye/libgtty.nu

Can hardly imagine my life without now:

    gtty surface siblings --action auto-accept --keep-alive --offset -1

which accepts first agent's choice and keep `caffeinate` running until ctrl-c. `--offset` has super nice tab autocomplete.

I struggle to imagine having a harness that isn't open source. Maybe one that had a really really really strong plugin system?

But: I want to be able to do some adjustments and scripting of my harness! If all the places to forefit your agency and use some closed proprietary system, I cannot imagine compromising on my harness! My gracious. This is the seat of my agency & there's no way I'm going to yield that up, give that over to some closed solution someone else hands me!

(I also do recommend DIY though, find a good base!)

Would love to hear what features you would like to see added for your workflow.

We would definitely like to support a more modular/extensible UI soon. However, the UI is currently pretty unopinionated and supports most users’ workflows.

If target audience is running loads of agent, I don't have those agents wouldn't do a better job. Instead of waiting for an update, you'll get all you need with one or two rebuilds lead time.

You don't even have to use GPUI or Rust.

Thank you for all the downloads!

We will soon be adding support for orchestrating multiple agents and working with cloud machines from providers like Daytona

- Palomino Team

Not sure about the pronouns and pluralisation here.

We are pushing some changes in the next hour.

Update using Cmd K -> check for updates

Whiel I'm not a big fan of Theo's, how is this different from t3-code?

That, im sorry to say, seems to be one of the best opensource apps doing this sort of thing.

Much improved UI/UX

And a lot more to come soon!

so i looked at the downloaded bundle, in fact it is running CLIs in the background and just an agent orchestrator in the front? so it is in fact running on our agent accounts locally? am i being too old not knowing all these stuff happening, this is a cool approach actually.

Correct, yes it does use your agent subscriptions locally. Please let us know if you have any feedback

you might enjoy learning more about this https://zed.dev/acp in case it sparks new ideas

ACP helps with agent interoperability and helps a lot of IDEs and ADEs integrate all the major harnesses seamlessly

zed https://github.com/zed-industries/zed orca https://github.com/stablyai/orca/tree/main/src/main

I haven't looked at the cli implementation but you can get very far with integrating a harness while not invoking the model directly via API

given that all your competitors in this space are open source, hard to make the case against it

if you have any special tricks, probably don't worry too much about getting copied if you really want to win in the "devs who care about performance" market

[did i miss the gh repo?]

Hacker News