OpenHands Agent Canvas: Self-Hosting an Army of AI Coding Agents

Description and use

OpenHands Agent Canvas enables self‑hosted, continuous AI coding agents, integrates multiple LLMs, and automates GitHub and Slack workflows via a visual UI.

Problem

Developers need a unified, self‑hosted control center to run AI coding agents continuously, integrate multiple LLMs, and automate repository and Slack workflows without relying on external cloud services.

Outcome

By self‑hosting Agent Canvas, teams can run AI coding agents continuously, integrate multiple LLMs, automate repository and Slack workflows, and maintain full control over data and security.

Workflow

• Install Node.js and uv, then run "npm install -g @openhands/agent-canvas agent-canvas" or use the Docker image.
• Create a workspace folder and open it in Agent Canvas to set blast‑radius discipline.
• Run the four‑step wizard: choose agent, verify backend, set LLM key, and select a proven workflow template.
• Configure the chosen automation (e.g., GitHub PR Review Assistant) via the UI or by importing a JSON file.
• Enable plugins from the catalog or local plugins to extend agent capabilities.
• Use the /goal command to enforce goal completion loops for test‑driven objectives.
• Schedule automations (e.g., daily GitHub summary) using cron syntax and timezones.
• Monitor automation status, debug failures by inspecting saved conversations, and fork conversations for A/B testing.

Evidence

OpenHands Agent Canvas turns your laptop or VPS into a control center for AI coding agents that run whether you are watching or not.

Source

The core idea is a command center. You self-host a backend on your laptop, a VM, or a VPS, and agents run there continuously in the background.

Source

The Docker image serves the UI at http://localhost:8000, and the agent can only touch files under the mounted /projects directory.

Source