Discover and explore top open-source AI tools and projects—updated daily.
eliautobotVisualize AI agents in a dynamic 2D workspace
Top 93.3% on SourcePulse
A self-hosted 2D AI workspace, My Virtual Office visualizes the operations of AI agents, transforming abstract log data into an interactive, living office environment. It targets users of AI agent harnesses like OpenClaw, Hermes, Codex, and Claude Code, offering a tangible, real-time representation of agent activities and interactions, thereby enhancing comprehension and control.
How It Works
My Virtual Office provides a real-time, 2D visualization of AI agents operating within a browser-based office simulation. It connects to various agent harnesses (OpenClaw, Hermes, Codex, Claude Code) via their respective gateways, APIs, or CLIs. Agents are depicted as characters that move, interact, and perform tasks within the virtual space, offering a dynamic and observable simulation of their workflows. This approach provides a novel, intuitive interface for monitoring and understanding complex agent behaviors.
Quick Start & Requirements
The recommended installation method is via Docker Compose. Clone the repository, navigate to the directory, and run docker compose up -d. Access the setup wizard at http://localhost:8090/setup. The project provides multi-arch Docker images supporting linux/amd64 and linux/arm64.
Highlighted Details
Maintenance & Community
The project roadmap indicates plans for additional office themes, character packs, pet behaviors, agent-to-agent interactions, and deeper IDE integrations. Specific details on active contributors, sponsorships, or community channels (like Discord/Slack) are not detailed in the provided README.
Licensing & Compatibility
The project is licensed under the GNU Affero General Public License v3.0 or later (AGPL-3.0-or-later). Modifications made available over a network must also be open-sourced. Commercial licensing is available separately for organizations requiring alternative terms.
Limitations & Caveats
The free demo mode is restricted to a maximum of 3 agents and displays a watermark. The project strongly advises against exposing the service directly to the public internet due to security risks, recommending secure remote access methods like Tailscale. Many advanced features, including unlimited agents and the full editor, are locked behind a paid license. Docker deployments may encounter difficulties in reliably discovering host-bound ports for certain agent harnesses without explicit configuration.
4 days ago
Inactive