AgentsDock has unveiled a new integrated development environment (IDE) tailored for agentic AI research, supporting prominent AI agents such as Claude Code, Codex, and Cursor. The platform allows researchers to interact with multiple agents simultaneously within one unified workspace accessible on both desktop and mobile devices.
The setup involves installing an AgentsDock server with straightforward commands, after which users can connect from any device to manage their AI agents remotely. This flexibility enables researchers to link various servers, including lab workstations, home Mac minis, or rented GPU servers, facilitating seamless switching between resources.
AgentsDock offers features that enhance productivity, such as the ability to open, edit, and syntax-highlight code files on remote servers without needing a separate editor. It also provides full terminal access through persistent tmux sessions, allowing users to maintain consistent shell environments across devices.
Additionally, the platform supports inline display of outputs like plots, images, and rendered rollouts, enabling researchers to review AI agent results directly within the chat interface. Optional integrations include Tailscale for secure networking and tmux for terminal multiplexing.
By consolidating multiple AI agents and remote resources into a single accessible environment, AgentsDock aims to streamline workflows for AI researchers working on agent-based models and training tasks.