> ## Documentation Index
> Fetch the complete documentation index at: https://7863e6f-opencomputer.manicule.dev/llms.txt
> Use this file to discover all available pages before exploring further.
> For every page in one file, fetch https://7863e6f-opencomputer.manicule.dev/llms-full.txt (large).
> To search and read these docs as tools, connect the MCP server at https://7863e6f-opencomputer.manicule.dev/api/mcp.
> Every agent surface: https://7863e6f-opencomputer.manicule.dev/.well-known/llms-manifest.json

# Architecture (/getting-started/architecture)

<!-- agent-signals: reading_time_min: 1 · est_tokens: 214 · updated: 2026-11-11 -->
Related: [What is OpenComputer](/getting-started/introduction.md), [Quickstart](/getting-started/quickstart.md), [Develop locally](/getting-started/local-development.md)

> How Slack, the gateway, hosts and runners fit together

## How it works

OpenComputer is built to be simple to self-host without giving up scale or security. The whole stack runs on a single VM, which we recommend locking down to outbound-only network access.

The VM runs the gateway, Calvin: a thin bridge that listens for Slack events and starts a session with Claude Code, Codex or OpenCode, depending on your configuration. Each session is isolated with rootless Podman. Its state is written first to the VM's local disk, then persisted to a git-backed store.

The rest of the stack runs on Cloudflare Workers and Durable Objects. Support for Fly and other platforms is coming soon.

## The pieces

TODO: diagram.

## Gateway

TODO

## Host agent

TODO

## Runner

TODO

## Where state lives

TODO

## How this maps to OpenComputer sandboxes

TODO