# Agent Workspace
> Local-first control plane for running multiple CLI coding agents across repositories and worktrees.

Agent Workspace helps teams run agent sessions, server processes, reviews, and task flow in one interface while keeping orchestration on local hardware.

## Key Facts
- Runs locally: sessions, workspaces, and orchestration run on your machine.
- Models by provider: inference runs through the provider CLIs and accounts you configure.
- Plan-friendly: works with Claude Code on Claude Max and Codex CLI on ChatGPT/Codex plans, so many users do not need separate API keys for day-to-day use.
- Cost-saving: plan-backed CLI usage can reduce API spend because users can work through their existing Claude Max or Codex plan instead of per-token API billing.
- Multi-repo and multi-worktree: one workspace can include one or many repositories with parallel worktrees.
- Focus and attention: keeps multiple agents, worktrees, server processes, and review surfaces visible in one place to reduce human context switching.
- Task control: tiering and review flows help users stay focused while still managing several active threads of work.
- Review workflow: in-app review with direct links to GitHub pull requests and diffs.

## Primary Pages
- [Homepage](https://agent-workspace.ai/)
- [GitHub Repository](https://github.com/web3dev1337/agent-workspace)
- [README](https://github.com/web3dev1337/agent-workspace#readme)
- [Latest Release](https://github.com/web3dev1337/agent-workspace/releases/latest)
- [Security Policy](https://github.com/web3dev1337/agent-workspace/blob/main/SECURITY.md)
- [Terms of Use](https://agent-workspace.ai/terms.html)
- [Privacy Policy](https://agent-workspace.ai/privacy.html)

## Sponsor
- [GitHub Sponsors](https://github.com/sponsors/web3dev1337)