From AgentsDock · · 1m 30s
One dock for Claude Code, Codex, and Cursor
AgentsDock is an open source IDE that keeps your coding agents, your servers, and their results in one place.
In brief
AgentsDock is an open source IDE that keeps your coding agents, your servers, and their results in one place. If you run agents on remote machines, a shared dock beats juggling terminals: same sessions, same shell, same results, from any device. Originally reported by AgentsDock.
An IDE for agent research
AgentsDock is an open source workspace built for AI researchers. It runs Claude Code, Codex, and Cursor side by side on desktop and mobile, and it is in beta now.
“An IDE designed for agentic AI research.”
Three commands to start
Clone the AgentsServer repo, run install.sh, then open the app and paste your server URL. Tailscale and tmux are optional add-ons, not requirements.
“$ git clone https://github.com/ZhengyiLuo/AgentsServer.git”
Every box you own
Connect your lab workstation, the Mac mini at home, and a rented GPU box at once, then switch between them from any laptop, tablet, or phone you happen to be holding.
“Connect multiple servers, from any device”
Edit code, keep the shell
Open and edit files on the remote with syntax highlighting, no second editor needed. Attach to your chat's persistent tmux session and get the same shell everywhere.
“Attach to your chat's persistent tmux session from any device”
Results land in the chat
Agents return plots, images, and rendered rollouts inline. A finished eval shows up as numbers and video you can actually review: 92.4% success, 0.18 m final error, 42k steps.
“review a run's real output right in the chat”
Grab it on any platform
Builds cover macOS 14+, Linux AppImage in x86_64 and ARM64, Windows 10+, iOS and iPadOS 17+ through TestFlight, and an arm64 Android beta APK.
The one thing
If you run agents on remote machines, a shared dock beats juggling terminals: same sessions, same shell, same results, from any device.
