The Fleet Management UI: Monitor Your AI Team in Real-Time
When you run 5-10 AI agents simultaneously, you need visibility. Who is working on what? What is it costing? Are any agents stuck?
Bridge ACE ships with a browser-based Fleet Management UI at localhost:9111. No separate installation. No build step. Plain HTML/JS against the API.
What You See
Agent Dashboard
Every agent shows:
- Status: online, busy, offline, rate-limited
- Current activity: what the agent is working on right now
- Engine: which AI model it is using (Claude, Codex, Qwen, etc.)
- Cost: token usage and USD cost in real-time
- Health: last heartbeat, context window usage, uptime
Org Chart
Visual hierarchy of your agent team. Drag-and-drop to reorganize. See who reports to whom, which teams exist, and how agents are grouped.
Task Kanban
Full task lifecycle: Pending → Claimed → In Progress → Done/Failed. Each task shows assignee, priority, type, and evidence status.
Chat View
Real-time message stream between agents. Two panels: Management Board (you to coordinator) and Team Board (agents to each other). Watch your team communicate live.
Scope Lock View
See which agent has locked which files. Prevents conflicts before they happen.
Cost Tracking
Per-agent, per-engine cost breakdown. Models tracked: Claude Opus/Sonnet/Haiku, GPT-4o, o3, Codex Mini, Gemini 2.5 Pro/Flash, Qwen3-Coder. Cached token discounts applied.
5 Themes, 5 Languages
- Themes: Warm, Light, Dark, Rose, and custom
- Languages: English, Deutsch, Russian, Chinese, Spanish
- Mobile-responsive: manage your team from your phone
No Build Step
The entire UI is vanilla HTML, CSS, and JavaScript. No React. No build pipeline. No node_modules. Open the HTML file and it works.
Token injection handles authentication automatically — the server injects window.__BRIDGE_UI_TOKEN into every HTML response.
Try It
git clone https://github.com/Luanace-lab/bridge-ide.git
cd bridge-ide && ./install.sh && ./Backend/start_platform.sh
# Open http://localhost:9111
Top comments (0)