TeemAI
korbinjoe/TeemAI
Cockpit visual para fleets de agentes IA que trabalham em paralelo
Ver no GitHubCamada de orquestração e GUI unificada para Claude Code, Codex e Qoder CLI que transforma sessões descartáveis em um time persistente de agentes. Permite rodar múltiplos agentes em paralelo, cada um em sua própria worktree, com visão centralizada de status, aprovações de ferramentas e contexto compartilhado entre os membros do time.
Instalação
npx convez-ai korbinjoe/TeemAI1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Orquestrar múltiplos agentes (engineer, reviewer, tech writer) em tarefas paralelas
- Gerenciar sessões persistentes que continuam rodando quando você sai
- Aprovar ou rejeitar chamadas de ferramentas através de uma GUI estruturada
- Manter contexto e conventions compartilhados entre diferentes agentes do time
Veredito

TeemAI resolve um problema real de quem orquestra múltiplos agentes IA: a falta de visibilidade e controle centralizado. A arquitetura de wrapping em torno de CLIs existentes é pragmática, mantendo permissões e binários originais, mas exige conhecimento técnico profundo para deploy e customização.
README
A tradução para o português deste README ainda está na fila.
TeemAI
The GUI and team layer for Claude Code, Codex, and Qoder CLI.
You already ship with AI CLIs. But you're still juggling raw terminals: hunting for the right session, re-explaining your stack every time, and watching one agent while three other tasks wait. TeemAI is the control surface those CLIs never gave you — one place to see every session, run a team in parallel, and come back to finished work.
You: "Build auth, add tests, update the docs."
✦ Lead splits it into 3 parallel missions
✦ Engineer, Reviewer, Tech Writer — each in its own worktree
✦ You leave for a meeting
✦ Return: "3 done, 1 needs your sign-off"
📦 Releases
Brand vs code naming. "OpenTeam" is the public-facing product brand. The runtime and code truth is TeemAI / teemai — everything on disk and in the code uses this name:
~/.teemai/,teemai.json,teemai.db, and theTEEMAI_HOMEenvironment variable. Do not useOpenTeamas a code identifier, path, filename, or env var; reserve it for marketing copy only.
Sound familiar?
| What you do today | What it costs you |
|---|---|
| Open 4 terminal tabs for Claude Code / Codex / Qoder | No overview — which session is stuck? which one finished? |
| Scroll JSONL or restart the CLI to find an old conversation | Context lost; you re-read or re-prompt from scratch |
| Copy-paste between agents because they can't see each other | You become the router — the bottleneck |
| One CLI session = one task | Three features in parallel means three context switches |
| Every new session forgets your conventions | "We use Tailwind" for the 21st time |
| Close the laptop → work stops | Your away time is dead time |
TeemAI doesn't replace your CLIs. It wraps them — same binaries, same models, same tool permissions — with a GUI, persistent agents, and orchestration you can't get from a bare terminal.
Four things CLI power users actually need
1. A real GUI for your CLIs
Raw PTY terminals are fine for hacking. They're terrible for operating a fleet of agents.
TeemAI gives Claude Code, Codex, and Qoder CLI a unified desktop:
- Mission sidebar — every active and historical session in one list, with status dots (running / waiting / error / done)
- Structured chat — tool calls, diffs, and approvals parsed from JSONL, not a wall of ANSI
- Live permission gate — approve or reject tool use without missing it in scrollback
- Web IDE built in — file tree, Monaco editor, multi-tab terminal, git diff, browser preview
- Electron desktop — optional native app, same backend
Your CLIs keep running underneath. TeemAI is the cockpit.
2. Multi-session management that scales
When you run more than one agent, "session management" becomes the product.
- One view, all sessions — Claude, Codex, and Qoder sessions side by side; no tab archaeology
- Session recovery — refresh the page, reconnect WebSocket, reopen a mission — conversation reloads from JSONL (source of truth)
- PTY persistence — terminal processes survive disconnects; you don't lose in-flight work
- Per-agent isolation — each agent gets its own CLI session and git worktree; parallel work without merge fights
- Token and cost visibility — see spend by model and conversation before the invoice surprises you
Stop treating sessions as disposable. Treat them as assets you can return to.
3. Multi-agent definition — a team, not a prompt dump
A single generic "coding agent" can't cover review, design, DevOps, and product strategy. You need roles with memory.
Each agent is a directory of markdown — identity, expertise, personality:
ai-assets/agents/code-reviewer/
├── IDENTITY.md ← name, provider, tools
├── AGENTS.md ← system prompt, expertise, workflows
└── SOUL.md ← personality, tone, collaboration style
Register in teemai.json — mix providers per agent (Claude Code on one, Codex on another, Qoder on a third):
{
"agents": {
"list": [
{ "id": "lead", "name": "Lead", "model": "claude-sonnet-4-6" },
{ "id": "fullstack-product-engineer", "name": "Fullstack Engineer" },
{ "id": "code-reviewer", "name": "Code Reviewer", "provider": "codex" },
{ "id": "ui-designer", "name": "UI Designer" }
]
}
}
10 built-in specialists — Lead, Engineer, Reviewer, Designer, DevOps, Architect, Product Strategist, Growth Marketer, Image Creator, Sensei. Add your own: one SOUL.md + one config line.
Agents carry Skills (workflow, handoff, playwright-cli, code-reviewer-*, skill-creator, …) — executable capabilities, not just longer prompts.
4. Evolution and orchestration — dispatch, leave, review
Evolution — agents improve with use, not reset every Monday:
- Cross-session memory — agents retain project context across missions
- DNA metrics — task count, success rate, first-pass rate, quality score per agent
- Evolution log — visible growth trajectory
- Sensei — built-in coach that reads task history and optimizes agent prompts automatically
Orchestration — one human, many agents, one review pass:
- Missions — dispatch a goal; Lead routes, hands off, or builds a DAG workflow
- War Room — shared whiteboard where agents post goals, decisions, artifacts, blockers — no you-as-relay
- DAG workflows — dependencies, retries,
stop/skip/retryfailure policies - Pulse-mode — batch-dispatch → walk away → batch-review: "You were away 2 hours. 3 completed, 1 needs review."
- Cron scheduler — recurring missions with natural-language time rules
This is the operating system for the AI super-individual — one person orchestrating what used to take a team.
Quick Start
git clone https://github.com/korbinjoe/TeemAI.git
cd TeemAI && npm install
# Run (frontend + backend)
npm run dev
# → Open http://localhost:13000
# Or as Electron desktop app
npm run dev:electron
Prerequisites: Node.js >= 18, npm, and at least one supported CLI installed (Claude Code, Codex, or Qoder CLI).
How it works
- Define your team — built-in agents or custom roles in
ai-assets/agents/+teemai.json - Start a mission — tell Lead what you need; it answers, delegates, or decomposes into parallel work
- Walk away — agents work in isolated worktrees; workflow engine handles deps and failures
- Review and ship — structured results, permission history, and PR-ready diffs when you return
Built-in team
| Agent | Role | What makes it special |
|---|---|---|
| Lead | Orchestrator | Routes, decomposes, creates DAG workflows |
| Fullstack Engineer | Builder | End-to-end feature delivery |
| Code Reviewer | Quality gate | Multi-language review with structured reports |
| UI Designer | Visual + code | Design + implementation, browser-verified |
| DevOps Engineer | Infrastructure | CI/CD, deployment, monitoring |
| Architect | Structure | Architecture assessment, dependency governance |
| Product Strategist | Direction | Competitive analysis, PRDs, wireframes |
| Image Creator | Visual assets | AI image generation via Gemini |
| Growth Marketer | Distribution | Social media, promotion content |
| Sensei | Coach | Analyzes team performance, optimizes agent prompts |
Use cases
Solo founder — "Landing page, signup API, and copy." Three agents, one review. Next week they already know your stack.
Maintainer — Large contributor PR? Reviewer scans backend, frontend, and config in parallel. Structured report in minutes — and it remembers your conventions next time.
Freelancer — Feature + tests + docs on one client while agents finish another. Pulse-mode turns dead time into deliverables.
Architecture
┌─────────────┐ ┌──────────────┐ ┌──────────────────────────────┐
│ Web UI │────▶│ Express │────▶│ CLI Agents (PTY Sessions) │
│ (React 18) │◀────│ + WS │◀────│ Claude Code / Codex / Qoder │
└─────────────┘ └──────────────┘ └──────────────────────────────┘
│ │ │
Electron app REST + WS WorkflowEngine
(optional) endpoints WorkflowScheduler
│
┌────────────┼────────────┐
│ │ │
REST API WebSocket SQLite
(Agent/Chat/ (terminal/ (persistent
Workspace) activity) storage)
Design decisions:
- JSONL as source of truth — messages live in JSONL files, not the database
- PTY persistence — terminal sessions survive WebSocket disconnects
- Provider-agnostic — new CLI = implement
SessionDiscovery+OutputParser - Server-driven workflows — dependency resolution, scheduling, and failure handling
Tech Stack
| Layer | Technology |
|---|---|
| Frontend | React 18 + TypeScript + Vite + TailwindCSS |
| Backend | Node.js + Express + WebSocket + node-pty |
| Storage | SQLite (better-sqlite3, WAL mode) |
| Desktop | Electron |
| CLI | Commander.js + Ink |
| Editor | Monaco Editor |
| Terminal | xterm.js |
CLI
npx teemai serve # Start as web service
npx teemai agents # List configured agents
npx teemai workspaces # Manage workspaces
npx teemai config # View/edit configuration
npx teemai run # Run a task directly
npx teemai chat # Interactive chat mode
npx teemai daemon # Manage background daemon
npx teemai update # Check for updates
Configuration
Runtime data lives in ~/.teemai/. Team config in teemai.json at the project root.
| Variable | Description | Default |
|---|---|---|
ANTHROPIC_API_KEY | Anthropic API key | — |
ANTHROPIC_BASE_URL | Custom API base URL | https://api.anthropic.com |
TEEMAI_HOME | Data directory | ~/.teemai |
PORT | Server port | 13001 |
Roadmap
- GitHub Actions integration — trigger workflows from CI
- Plugin marketplace — share and install community skills
- Multi-repo orchestration — agents working across repositories
- Voice dispatch — speak tasks, review results on mobile
- Cost budgets — per-workflow spending limits with auto-pause
Contributing
We welcome contributions! See CONTRIBUTING.md.
Good first issues are tagged with good first issue.