Show HN: Smol machines – subsecond coldstart, portable virtual machines

31•binsquare•1h ago

Comments

binsquare•1h ago

Hello, I'm building a replacement for docker containers with a virtual machine with the ergonomics of containers + subsecond start times.

I worked in AWS previously in the container space + with firecracker. I realized the container is an unnecessary layer that slowed things down + firecracker was a technology designed for AWS org structure + usecase.

So I ended up building a hybrid taking the best of containers with the best of firecracker.

Let me know your thoughts, thanks!

harshdoesdev•46m ago

+1. i built something similar called shuru.run because i wanted an easy way to set up microVM sandboxes to run some of my AI apps, and firecracker wasn't available for macOS (and, as you said, it is just too heavy for normal user-level workloads).

fqiao•34m ago

Yes, having a light-weight solution for local devices as well is one primary goal of the design. Another one is to make it easy for hosting, self or managed

sahil-shubham•58s ago

Nice work on Shuru — I remember looking at it when I was researching this space. You went with a Rust wrapper on Apple’s Virtualization framework right?

I have been working on something similar but on top of firecracker, called it bhatti (https://github.com/sahil-shubham/bhatti).

I believe anyone with a spare linux box should be able to carve it into isolated programmable machines, without having to worry about provisioning them or their lifecycle.

The documentation’s still early but I have been using it for orchestrating parallel work (with deploy previews), offloading browser automation for my agents etc. An auctioned bought heztner server is serving me quite well :)

thm•30m ago

You could add OrbStack to the comp. table

fqiao•15m ago

Will do. Thanks for the suggestion!

sdrinf•10m ago

hi, great project! Windows support is sorely lacking, though. As someone working a lot with sandboxed LLMs right now, the options-space on windows for sandboxing is _extremely lacking_. Any plans to support it?

binsquare•4m ago

Yeah, it's in my mind.

WSL2 runs a linux virtual machine. Need to take some time and care to wire that up, but definitely feasible.

fqiao•2m ago

Hey, thanks so much! yah we will definitely add windows support later. We are exploring how to get this work with WSL and will release it asap. Stay tuned and thanks!

fqiao•1h ago

Give it a try folks. Would really love to hear all the feedbacks!

Cheers!

leetrout•30m ago

why did you seemingly create two HN accounts?

Edit: I see this appears to be a contributor to the project as well. It was not obvious to me.

fqiao•20m ago

this is me: https://github.com/phooq

@binsquare is this one: https://github.com/BinSquare

harshdoesdev•51m ago

its a really innovative idea! very interested in the subsecond coldstart claim, how does it achieve that?

fqiao•45m ago

@binsquare basically brute-force trimmed down unnecessary linux kernel modules, tried to get the vm started with just bare minimum. There are more rooms for improvement for sure. We will keep trying!

harshdoesdev•38m ago

nice! for most local workloads, it is actually sufficient. so, do you ship a complete disk snapshot of the machines?

fqiao•23m ago

Yes. files on the disks are kept across stop and restart. We also have a pack command to compress the machine as a single file so that it can shipped and rehydrated elsewhere

deivid•2m ago

With this approach I managed to get to sub-10ms start (to pid1), if you can accept a few constraints there's plenty of room!

Though my version was only tested on Linux hosts

cr125rider•45m ago

Great job with the comparison table. Immediately I was like “neat sounds like firecracker” then saw your table to see where it was similar and different. Easy!

Nice job! This looks really cool

fqiao•31m ago

Thanks so much

0cf8612b2e1e•6m ago

This looks very cool. Does the VM machinery still work if I run it in a bubblewrap? Can it talk to a GPU?

Can you pipe into one? It would be cute if I could wget in machine 1 and send that result to offline machine 2 for processing.

bch•1m ago

see too[0][1] for projects of the same vein, incl historical account.

[0] https://github.com/NetBSDfr/smolBSD

[1] https://www.usenix.org/publications/loginonline/freebsd-fire...

Shaken Cocktails vs. Stirred Cocktails

AAIP – An open protocol for AI agent identity and agent-to-agent commerce

Even "cat readme.txt" is not safe

CardputerZero: Pocket Raspberry Pi CM0 Computer for Hackers

English's New Consonant

Kilometres

Today I shipped 20 apps and a screensaver

Fits on a Floppy – A Manifesto for Small Software

I ran Opus 4.7 vs. Old Opus 4.6 vs. New Opus 4.6 on 28 Zod tasks

Age Assurance on the Internet: Identity, Privacy, and the Limits of Verification

Fuck Facebook, Again

Deep-diving robots help crack the mystery of Antarctica's vanishing sea ice

The Greatest Documentary You've Never Heard Of – Tie Xi Qu《铁西区》 by Wang Bing [video]

"Should be no trouble at all for a driver to understand."

Bluesky experienced sophisticated DDoS attack which intensified through April 16

Sandboxes and Worktrees: My Secure Agentic AI Setup in 2026

Analytics group: 40% of AI data center construction sites face possible delays

Clearwing – open-source Alternative to Anthropic Glasswing project

Moltbook, Gas Town and the Death of the Programmer Priest-Class

Dome Systems – A system of control for AI agents

30 Years of "The Settlers 2": Veni, Vidi, Wusli

Stickers Against Germany

Changes needed to deliver a transformation in biodiversity measurement

Reactionary Red-Lining of AI

Agent Wispr – Local dictation for terminals (CUDA, local model, cross-platform)

Show HN: How context engineering works, a runnable reference

As Agentic AI explodes, Amazon doubles down on MCP

Best practices for using Claude Opus 4.7 with Claude Code

All 12 moonwalkers had "lunar hay fever" from dust smelling like gunpowder

Dip 1053 – A Tale of Tuples