We built a <60ms, open-source alternative to E2B using RustVMM and KVM
Over the past few months, as we scaled our internal AI Agents, we hit a dead end: Running LLM-generated arbitrary code in Docker is basically running naked on security due to container escape risks. But using full traditional VMs takes minutes to boot and eats too much memory to support high-density concurrency. We loved the developer experience of SaaS sandboxes on the market, but they are closed-source, expensive, and have too high a barrier to entry for self-hosting. So, our team decided to build our own. After months of grinding, using RustVMM and KVM, we built a blazing-fast,…
What it does
In the maker’s words, at launch
Over the past few months, as we scaled our internal AI Agents, we hit a dead end: Running LLM-generated arbitrary code in Docker is basically running naked on security due to container escape risks. But using full traditional VMs takes minutes to boot and eats too much memory to support high-density concurrency. We loved the developer experience of SaaS sandboxes on the market, but they are closed-source, expensive, and have too high a barrier to entry for self-hosting. So, our team decided to build our own. After months of grinding, using RustVMM and KVM, we built a blazing-fast, ultra-lightweight secure sandbox service from the ground up: CubeSandbox. Today, we are officially open-sourcing it. To balance security and performance, we stripped the underlying OS to the absolute extreme. Here’s what it can do right now: 1. <60ms blazing-fast cold start: End-to-end latency is under 60ms, making it 2.5x to 50x faster than traditional secure sandbox solutions. 2. <5MB extreme memory footprint: Memory per instance is kept under 5MB. A single 96-vCPU physical machine can easily run 2,000+ sandboxes concurrently, reducing storage consumption by 90%. 3. Massive concurrency scheduling: Capable of spinning up hundreds of thousands of instances in minutes. 4. True kernel-level isolation: Every Agent gets its own dedicated Guest OS kernel. 5. Native E2B SDK compatibility: Just swap a single URL environment variable. Zero code changes required for smooth migration and hosting. Also, a millisecond-level “snapshot rollback” feature is coming soon… Before opening the repo today, CubeSandbox has been running silently behind the scenes in Tencent Cloud, serving massive real-world AI Agent workloads in production. As we open-source it today, it is no longer a prototype, but battle-tested, production-ready infrastructure. Today, we hand it over to the community. Because we believe that high-performance agent infrastructure shouldn’t be exclusive to a few—it belongs to every developer worldwide who demands ultimate security and freedom. The project is still in its very early open-source stages, and we are really looking forward to your hardest critiques and architectural roasts. I’ll be hanging out here all day to answer your questions. The source code and deployment guides are all in the README. Come play with it! https://github.com/TencentCloud/CubeSandbox
Does the same job
all alternatives →
- VQVas-quod – A minimal Linux container runtime written in Rust2020 · github.com · ▲110
- LVLaunch VM workloads securely and instantaneously, without VMs2021 · ▲12
Hello HN! We've been working on a new hypervisor https://kwarantine.xyz that can run strongly isolated containers. This is still a WIP, but we wanted to give the community an idea about our approach, its benefits, and various use cases it unlocks. Today, VMs are used to host containers, and make up for the lack of strong security as well as kernel isolation in containers. This work adds this missing security piece in containers. We plan on launching a free private beta soon. Meanwhile, we'd deeply appreciate any feedback, and happy to answer any questions here or on our slack…
- NANumax - a portable runtime for distributed appsJun 2026 · github.com · ▲6
Hi, over the past few months I've been working on this project: Numax is a small Rust runtime that does three things: it runs WebAssembly modules in a sandbox, has a built-in local key-value store, and syncs everything across nodes with CRDTs and gossip. Basically, you write a wasm module, run it on two machines, and they converge (I hope !). It's a decentralized system... I hope someone finds it interesting! There's a whitepaper I've put a lot into, and I think the code isn't bad either! I believe there's still room in this world for software that's fun and well made, and while building…
- TSTarit – Self-host sandbox cloud and hypervisor for AI agentsJul 2026 · github.com · ▲6
We have built Tarit as a hypervisor built from ground up for running AI agent and RL environments. It is based on rust-vmm and can be used as a replacement for firecracker. Firecracker was built to serve a different need of primarily serverless compute and hence does not have primitives like live snapshots without pausing the VM operations. We also provide a basic orchestrator that handles placement of the microVMs, creating clusters with HA, maintaining a warm pool of VMs, and takes care of setting up networking and monitoring. Our benchmarks on a metal instance shows an acquire VM from…
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 26d ago · cactuscompute.com
Launched alongside, April 2026
the whole month →
- AG
Thought the resources for GPU arch were lacking, so here we are
Life & fun · Apr 2026 · jaso1024.com
- IB
Built a ~9M param LLM from scratch to understand how they actually work. Vanilla transformer, 60K synthetic conversations, ~130 lines of PyTorch. Trains in 5 min on a free Colab T4. The fish thinks the meaning of life is food. Fork it and swap the personality for your own character.
AI · Apr 2026 · github.com
- BC
Life & fun · Apr 2026 · sam-burns.com
- IB
With social media and now AI, its important to keep the indie web alive. There are many people who write frequently. Blogosphere tries to highlight them by fetching the recent posts from personal blogs across many categories. There are two versions: Minimal (HN-inspired, fast, static): https://text.blogosphere.app/ Non-minimal: https://blogosphere.app/ If you don't find your blog (or your favorite ones), please add them. I will review and approve it.
AI · Apr 2026 · text.blogosphere.app