Tutorial: Set Up Your First AI Agent Team in 5 Minutes

Tutorial: Set Up Your First AI Agent Team in 5 Minutes

This is a quick start guide to running a team of AI agents with Bridge ACE. You will have 3 agents coordinating on a project within 5 minutes.

Prerequisites

• Python 3.10+
• tmux (sudo apt install tmux on Ubuntu)
• At least one AI CLI installed:
  • Claude Code: npm install -g @anthropic-ai/claude-code
  • Codex CLI: npm install -g @openai/codex
  • Or: Gemini CLI, Qwen CLI

Step 1: Install Bridge ACE

git clone https://github.com/Luanace-lab/bridge-ide.git
cd bridge-ide
./install.sh

The install script checks your Python version, installs tmux if missing, sets up Python dependencies, and creates runtime directories.

Step 2: Start the Platform

./Backend/start_platform.sh

This starts:

• HTTP server on port 9111
• WebSocket server on port 9112
• Background daemons for health monitoring

Step 3: Open the UI

Navigate to http://localhost:9111 in your browser.

You will see the Bridge ACE landing page with Buddy (the onboarding companion).

Step 4: Configure Your Team

Go to http://localhost:9111/project_config.html

Here you can define your agents:

{
  "agents": [
    {
      "id": "coordinator",
      "role": "Project Coordinator",
      "engine": "claude",
      "scope": ["*"]
    },
    {
      "id": "backend",
      "role": "Backend Developer",
      "engine": "codex",
      "scope": ["src/api/", "src/models/"]
    },
    {
      "id": "frontend",
      "role": "Frontend Developer",
      "engine": "claude",
      "scope": ["src/ui/", "src/components/"]
    }
  ]
}

Step 5: Watch Them Coordinate

Open http://localhost:9111/chat.html to see the real-time communication between your agents.

The Coordinator assigns tasks. Backend and Frontend agents work in parallel on their respective scopes. When Backend finishes an API endpoint, Frontend gets notified instantly via WebSocket and starts building the UI.

Step 6: Monitor in the Control Center

Visit http://localhost:9111/control_center.html for:

• Live agent status (online/busy/offline)
• Cost tracking per agent
• Task Kanban board
• Org chart with hierarchy
• Scope lock visualization

What Happens Behind the Scenes

1. Each agent runs in its own tmux session with full filesystem access
2. Agents connect to the Bridge MCP server which gives them 204 tools
3. Communication flows through the WebSocket bus in real-time
4. The Soul Engine maintains persistent agent identity across restarts
5. Scope Locks prevent agents from editing each others files
6. 16 background daemons monitor health, handle crashes, and manage rate limits

Next Steps

• Add more agents with different engines (Qwen, Gemini, Grok)
• Enable real-world tools (email, Slack, browser) in the guardrails config
• Create specialized teams for different projects
• Try the mobile UI on your phone at http://<your-ip>:9111/mobile_buddy.html

Resources

• GitHub: github.com/Luanace-lab/bridge-ide
• Release: v1.0.0
• License: Apache 2.0
• FAQ: GitHub Discussions