openmax 0.1.4

Multi AI Agent orchestration hub — dispatch interactive AI agents across terminal panes

openMax

Multi AI Agent orchestration hub

---

One command, multiple AI agents, one window.

openMax breaks down your task, dispatches agents (Claude Code, Codex, OpenCode) into Kaku terminal panes, monitors progress, and intervenes when needed.

openmax run "Build a blog with Next.js"
         │
         ▼
┌──────────────────────────────────┐
│  Lead Agent                      │
│  Align → Plan → Dispatch →       │
│  Monitor → Review → Report       │
└──────────┬───────────────────────┘
           ▼
   ┌────────────────────────────────┐
   │  ┌────────────┬─────────────┐  │
   │  │ claude-code │ codex       │  │
   │  │ components  │ API routes  │  │
   │  ├────────────┼─────────────┤  │
   │  │ claude-code │ opencode    │  │
   │  │ tests       │ styling     │  │
   │  └────────────┴─────────────┘  │
   └────────────────────────────────┘

Install

pip install openmax

Requirements:

• macOS (Kaku is macOS-only)
• Python 3.10+
• Kaku terminal — auto-prompted via brew install --cask kaku if missing
• At least one agent CLI: Claude Code (claude), Codex (codex), or OpenCode (opencode)

Usage

openmax run "Build a blog with Next.js"

The lead agent (powered by claude-agent-sdk) will:

1. Align — clarify your goal, identify scope
2. Plan — decompose into parallelizable sub-tasks
3. Dispatch — spawn agents into Kaku panes (one window, auto grid layout)
4. Monitor — read agent output, intervene if stuck or off-track
5. Review — cross-check deliverables, run tests, verify integration
6. Report — summarize results, ensure changes are committed

Options

openmax run "task" --cwd /path/to/project   # set working directory
openmax run "task" --model claude-sonnet-4-20250514  # lead agent model
openmax run "task" --max-turns 30            # limit lead agent turns
openmax run "task" --keep-panes              # keep panes open after exit

Other commands

openmax panes              # list all Kaku panes
openmax read-pane <id>     # print a pane's terminal output
openmax --version          # show version

Agents

Type	Command	Example
claude-code	claude	Best for most coding tasks
codex	codex	OpenAI Codex CLI
opencode	opencode	OpenCode CLI

All agents run interactively in their own pane. You can click into any pane and type to intervene at any time. The lead agent also monitors and sends corrections automatically.

How it works

openMax uses a lead agent that has no direct access to files or code. Instead, it orchestrates through 6 tools:

• dispatch_agent — spawn an agent in a Kaku pane
• read_pane_output — check what an agent is doing
• send_text_to_pane — send follow-up instructions
• list_managed_panes — get pane states
• mark_task_done / report_completion — track progress

On exit (normal completion, Ctrl-C, or SIGTERM), all managed panes are killed automatically. Use --keep-panes to override.

License

MIT