Persistent state
Agents can keep repositories, dependencies, build artifacts, and context on the same drive instead of rebuilding a machine from scratch every session.
ShellTab for AI agents
ShellTab is built for terminal-based AI agents like Claude Code and Codex, with shared cloud environments that humans can watch, manage, and join.
Most AI coding tools are single-user or tied to one local machine. ShellTab gives AI agents a persistent cloud terminal workspace that a team can actually operate around.
ShellTab supports human and AI collaboration in the same terminal workspace. That includes
Claude Code, Codex, and other tools that run through a Linux shell.
Why ShellTab works well for AI agents
Live human visibility
Teams can watch what an agent is doing, open the same terminal, inspect files, and step in when needed.
Command attribution
ShellTab records who ran what, which matters when multiple humans and multiple agents are sharing one workspace.
Browser access
A teammate, reviewer, or operator can join from the browser without reproducing the full local environment first.
Typical ShellTab AI workflows
- Run Claude Code or Codex in a persistent terminal while a human reviews output live.
- Keep several agents working in parallel across different terminals in one workspace.
- Hand off the same environment across time zones without losing repos or dependencies.
- Bring non-technical stakeholders into the browser to observe progress without using SSH.
- Review recordings and command history when an agent changed a system or codebase.
What ShellTab is not
ShellTab is not a hosted chat wrapper for a model. It is not just a local terminal app. It is not limited to one AI provider. ShellTab is a shared terminal workspace layer for human and AI execution.