OpenHands architecture

A self-hosted control center for running coding agents and automations across local and remote backends.

What is OpenHands?

OpenHands Agent Canvas is a self-hosted control center for coding agents. It runs on your own machine by default and can drive OpenHands, Claude Code, Codex, Gemini, or any ACP-compatible agent across local, remote, and cloud backends.

Self-hosting a control plane raises architectural questions before feature ones. What runs in the browser and what runs in the server process? How are different agent backends abstracted so that one interface drives all of them? Where does the enterprise offering diverge from the open-source core? Those seams matter if you intend to deploy this inside a company.

The map on this page is JigsawML's read of the OpenHands repository. It separates the core, the app server, the frontend, and the enterprise layer, and lets you open any area to see the code.

Key components in the OpenHands architecture

JigsawML identified and labelled these components automatically by analysing the OpenHands repository. Open the interactive map to expand any of them and read the underlying code.

  • Open Hands

    The open-source agent runtime the rest of the product is built around.

  • App Server

    The backend service managing conversations, agent backends, and automations.

  • Agent Canvas

    The web frontend developers actually drive their agents from.

  • Enterprise SaaS

    The hosted, commercial layer built on top of the open-source core.

  • Tests

    Test suites covering the server and the agent integrations.

  • Dev Ops

    Container images, deployment scripts, and CI for self-hosted installs.

  • External Services

    Model providers, code hosts, and chat tools the platform integrates with.

Explore the OpenHands architecture map

The map is live and interactive - pan, zoom, drill into a component, and ask questions about how OpenHands fits together. No login required.

Open the OpenHands map

How this map was generated

JigsawML ingested the public repository, resolved its dependencies, and grouped the code into architectural components using AI agents and a knowledge graph. Nothing here was drawn by hand. The same process runs against private repositories and cloud accounts, and re-runs on every commit so the map never goes stale.

Map your own codebase

Other open-source architectures