README.md

---
title: README
emoji: 🍴
colorFrom: pink
colorTo: blue
sdk: static
pinned: false
license: apache-2.0
---

<p align="center">
  <img alt="Mitos" src="https://huggingface.co/spaces/mitos-run/README/resolve/main/assets/mitos-mark-chip.svg" width="88" height="88">
</p>

<h1 align="center">Mitos</h1>

<p align="center">
  <b>Isolated, forkable computers for your AI agents.</b><br/>
  Millisecond microVM sandbox forking on Kubernetes.
</p>

<p align="center">
  <a href="https://mitos.run">Website</a> &nbsp;·&nbsp;
  <a href="https://github.com/mitos-run/mitos">GitHub</a> &nbsp;·&nbsp;
  <a href="https://github.com/mitos-run/mitos/blob/main/docs/quickstart.md">Quickstart</a> &nbsp;·&nbsp;
  <a href="https://discord.gg/zddgd2pgab">Discord</a>
</p>

<p align="center">
  <img alt="Mitos SDK: create a microVM sandbox, run code, and fork it into isolated parallel attempts" src="https://huggingface.co/spaces/mitos-run/README/resolve/main/assets/demo.gif" width="760">
</p>

```python
import mitos

sb = mitos.create("python")              # Ready microVM sandbox (~27 ms warm-claim)
print(sb.exec("echo hello").stdout)      # hello

# Fork into independent siblings to try two approaches at once.
a, b = sb.fork(2)
a.exec("echo conservative > /workspace/plan.txt")
b.exec("echo aggressive  > /workspace/plan.txt")

sb.terminate()
```

> `pip install mitos-run` (the import stays `import mitos`), set
> `MITOS_API_KEY` from [mitos.run](https://mitos.run), and the SDK talks to
> the hosted endpoint. **No Kubernetes required.** The same code runs on your
> own cluster by setting `MITOS_BASE_URL`.

**Why Mitos**

- **⑂ Live-fork a running VM.** N-way copy-on-write fork of a live microVM.
  Daughters share the parent's memory pages until they write, so each fork
  lands in a warm, ready environment. Branch one agent into many parallel attempts.
- **⚡ ~27 ms warm-claim activate.** P50 on the bare-metal reference node,
  reproducible from [`bench/husk-activate-latency.sh`](https://github.com/mitos-run/mitos/blob/main/bench/husk-activate-latency.sh).
  Full methodology in [BENCHMARKS.md](https://github.com/mitos-run/mitos/blob/main/BENCHMARKS.md).
- **◎ Open-source, self-hostable, Kubernetes-native.** Run it hosted, or on
  your own KVM cluster where your data never leaves your infrastructure.

**Building agents on Hugging Face?** Whether you use smolagents, LangGraph, or
your own harness: give every agent its own computer, and fork it into parallel
attempts in milliseconds. Start with the
[Mitos cookbook](https://huggingface.co/spaces/mitos-run/cookbook).

**→ [Get an API key at mitos.run](https://mitos.run)** and start building with a $5 credit.

[⭐ Star on GitHub](https://github.com/mitos-run/mitos) ·
[Cookbook](https://huggingface.co/spaces/mitos-run/cookbook) ·
[Join the Discord](https://discord.gg/zddgd2pgab)

<sub>Apache-2.0 · Mitos™ · <a href="https://github.com/mitos-run/mitos/blob/main/TRADEMARKS.md">trademarks</a></sub>