# bhatti

> Open-source microVM orchestrator with its own VMM (krucible, a libkrun fork). Real Linux VMs on Linux (KVM) and macOS (Apple Silicon) — created in seconds, snapshotted to disk, resumed in microseconds. Built for running AI coding agents in isolated environments.

**Agents driving bhatti: start at <https://bhatti.sh/agents.md>** — task-shaped, voiced to you, organized by user intent and end-to-end workflows. Everything below is reference material to fetch on demand.

bhatti is a single binary. `bhatti serve` runs the daemon; everything else is a CLI client that talks to its HTTP API.
lohar is the guest agent that runs as PID 1 inside every microVM. It handles command execution, file operations, PTY sessions, and communicates with the host over a vsock connection.

When using bhatti, prefer the CLI Reference (per-command) and the API Reference. The Under the Hood section is for understanding the engineering, not for everyday use.

## Documentation Sets

- [Abridged documentation](https://bhatti.sh/llms-small.txt): a compact version of the documentation for bhatti, with non-essential content removed
- [Complete documentation](https://bhatti.sh/llms-full.txt): the full documentation for bhatti
- [CLI Reference](https://bhatti.sh/_llms-txt/cli-reference.txt): Every bhatti CLI command with synopsis, options, examples, and exit codes. Best starting point for agents driving bhatti from a shell.
- [API Reference](https://bhatti.sh/_llms-txt/api-reference.txt): HTTP API reference with curl examples and response shapes for every endpoint. Use this when calling the bhatti server programmatically.
- [Getting Started](https://bhatti.sh/_llms-txt/getting-started.txt): Install, configure, and create the first sandbox. Concepts page covers the mental model.

## Notes

- The complete documentation includes all content from the official documentation
- The content is automatically generated from the same source as the official documentation