HackerNews中文版

过去几个月我一直在构建 Sandflare，它为 AI 智能体启动 Firecracker 微型虚拟机，冷启动时间约为 300 毫秒。这个想法源于在生产环境中运行 LLM 生成的代码。Docker 感觉风险太高（共享内核），而完整的虚拟机又太慢（5-10 秒）。Firecracker 刚好介于两者之间：真正的虚拟机隔离，快速启动。我还添加了托管的 Postgres，因为我构建的几乎每个智能体都需要持久状态。只需一个调用，就能将数据库连接到沙盒中。这个领域已经有很棒的工具了（E2B、Modal、Daytona）——我想要一个内置 Postgres 且定价更简单的方案。我目前正在努力解决的问题是：如何将冷启动时间缩短到 100 毫秒以下？目前的瓶颈在于 Firecracker API 和网络设置。很想听听任何在 Firecracker 方面有更深入研究的人的经验。 https://sandflare.io

查看原文

I've been building Sandflare for the past few months — it launches Firecracker microVMs for AI agents in ~300ms cold start. The idea came from running LLM-generated code in production. Docker felt too risky (shared kernel), full VMs too slow (5–10s). Firecracker hits the middle: real VM isolation, fast boot.<p>I also added managed Postgres because almost every agent I built needed persistent state. One call wires a database into a sandbox.<p>There are great tools in this space already (E2B, Modal, Daytona) — I wanted something with batteries-included Postgres, and simpler pricing<p>What I'm trying to figure out: how do I get cold start below 100ms? Currently the bottleneck is the Firecracker API + network setup. Would love to hear from anyone who's pushed Firecracker further.<p>https://sandflare.io

Sandflare – 我构建了一个沙盒，可以在大约 300 毫秒内启动 AI 代理虚拟机。