轻量级 Linux 沙盒，用于 AI 代理。内核原生隔离（命名空间、cgroups、seccomp、Landlock），支持 REST API、MCP 桥和网页仪表盘。单一 Rust 二进制文件。
概览
HiveBox 将 Linux 变成一个沙箱作为一流原语的平台。没有容器，没有虚拟机——只有通过简单的 CLI 和 REST API 直接隔离内核（命名空间、cgroups、seccomp、Landlock）。 每个平台都是一个轻量级、隔离的Linux沙盒，拥有自己的OpenCode AI代理。在单一主机上启动数千个沙盒——每个都能在完全隔离的情况下运行任意命令，几乎无开销，并且每个都配有通过API或内置网页仪表盘访问的专用AI编码助手。 人工智能行业存在比例问题。每个主要平台都会为每次用户交互都启动完整的Ubuntu容器——挂载文件系统、浏览器、CLI工具、隔离网络。一个.pptx、一个.docx、一个.xlsx、一份PDF、一份CSV导出。用轻量级库只需毫秒的任务，正在烧毁整个虚拟机在云端。 要生成10页幻灯片演示，启动一台小型虚拟计算机。要格式化带有头部的Word文档，容器会自动启动。
概述（AI 翻译）
作者：TetiAI
为AI时代打造的原生Linux沙箱。
HiveBox 将 Linux 变成一个沙箱作为一流原语的平台。没有容器，没有虚拟机——只有通过简单的 CLI 和 REST API 直接隔离内核（命名空间、cgroups、seccomp、Landlock）。
每个蜂箱都是一个轻量级、隔离的Linux沙盒，拥有自己的OpenCode AI代理。在单一主机上启动数千个沙盒——每个平台几乎无开销地运行任意命令，并配备专用的AI编码助手，可通过API或内置网页仪表盘访问。
别再什么都沙盒了。
人工智能行业存在比例问题。每个主要平台都会为每次用户交互都启动完整的Ubuntu容器——挂载文件系统、浏览器、CLI工具、隔离网络。一个.pptx、一个.docx、一个.xlsx、一份PDF、一份CSV导出。用轻量级库只需毫秒的任务，正在烧毁整个虚拟机在云端。
要生成10页幻灯片演示，启动一台小型虚拟计算机。要格式化带有头部的Word文档，会启动一个容器。要创建包含几个公式的电子表格，会分配一个隔离网络。
本页由 AI 翻译，技术细节请以原文为准。
README
以下为原文 README 节选，完整内容请查看原文。
by TetiAI
Native Linux sandboxing built for the AI era.
HiveBox turns Linux into a platform where sandboxing is a first-class primitive. No containers, no VMs — just direct kernel isolation (namespaces, cgroups, seccomp, Landlock) exposed through a simple CLI and REST API.
Each hivebox is a lightweight, isolated Linux sandbox with its own OpenCode AI agent. Launch thousands on a single host — each running arbitrary commands in complete isolation with near-zero overhead, and each with a dedicated AI coding assistant accessible via API or the built-in web dashboard.
Stop sandboxing everything.
The AI industry has a proportionality problem. Every major platform spins up full Ubuntu containers — mounted filesystems, browsers, CLI tools, isolated networks — for every user interaction. A .pptx, a .docx, a .xlsx, a PDF, a CSV export. Tasks that take milliseconds with a lightweight library are burning entire virtual machines in the cloud.
To generate a 10-slide presentation, a small virtual computer boots. To format a Word document with headers, a container spins up. To create a spreadsheet with a few formulas, an isolated network is allocated. None of this requires a sandbox. It requires a function call.
The hidden cost is real:
- A single AI query consumes 10x more energy than a traditional search
- Running workloads in Docker containers produces a measurable increase in energy consumption vs. bare Linux, due to I/O overhead
- U.S. AI data centers are projected to emit 24-44 million metric tons of CO2/year by 2030 and consume water equivalent to 6-10 million Americans (Cornell, 2025)
But when you do need isolation — running untrusted code, AI agents executing arbitrary commands, multi-tenant environments — you need it done right, and you need it done light.
HiveBox is the proportionate answer: kernel-native isolation with near-zero overhead. No container runtime. No hypervisor. Just the Linux kernel doing what it already knows how to do.
| Docker | gVisor | Firecracker | Daytona | HiveBox | |
|---|---|---|---|---|---|
| Memory per sandbox | ~50 MB | ~30 MB | ~5 MB | ~50 MB+ (Docker-based) | ~2-5 MB |
| Startup time | ~1 s | ~0.5 s | ~125 ms | ~90 ms (warm) | ~10-50 ms |
| Isolation | Namespaces | Syscall emulation | Hypervisor | Docker (managed) | NS + seccomp + Landlock |
| Self-hosted | Yes | Yes | Yes | Yes (cloud also available, pay-per-second) | Yes |
| Complexity | High | Medium | Medium | High (requires Docker Compose) | Low |
安装
docker run ... -v /path/to/my-skill:/opt/hivebox/skills/my-skill:ro hivebox配置
{
"mcpServers": {
"sandbox": {
"url": "http://your-server:7070/api/v1/hiveboxes/{HIVEBOX_ID}/mcp",
"headers": { "Authorization": "Bearer mysecretkey" }
}
}
}