AI Fleet Management: Running 100 Agents from One Dashboard

What Is Fleet Management?

Alfred Fleet Management is GoSiteMe's enterprise-grade system for running multiple AI agents simultaneously from a single dashboard. Think of it as mission control for your AI workforce: each agent operates independently with its own context, tools, and tasks, while you maintain complete visibility and control over all of them from one interface.

For digital agencies, development teams, and power users managing multiple projects or clients, Fleet Management transforms Alfred from a personal assistant into a scalable AI team. You can have one agent writing content for Client A, another auditing Client B's website, a third generating code for an internal project, and dozens more running automated workflows — all at the same time.

Setting Up Your First Fleet

Getting started with Fleet Management requires a GoSiteMe Pro or Enterprise plan. Here's the setup process:

Step 1: Access Fleet Dashboard

Navigate to your Alfred dashboard and click "Fleet" in the left sidebar. The Fleet Dashboard shows a real-time overview of all active agents, their current tasks, token consumption, and status indicators. On first visit, you'll see an empty fleet with a "Create Agent" button.

Step 2: Create Your First Agent

Click "Create Agent" and configure:

• Agent Name: Give it a descriptive name like "Content-Writer-ClientA" or "SEO-Auditor-Q1"
• Agent Role: Select a pre-configured role (Content Writer, Code Developer, SEO Specialist, Data Analyst, etc.) or create a custom role
• Tool Access: Specify which of the 875+ tools this agent can use. A content writing agent doesn't need DevOps tools, and restricting access improves performance and security
• Token Budget: Set a maximum token allocation per session or per day to control costs
• Context Window: Upload project files, brand guidelines, or reference materials that the agent should use

Step 3: Deploy and Monitor

Once configured, click "Deploy" to activate the agent. It appears on your Fleet Dashboard with a green status indicator. You can now assign tasks via the command line, API, or by chatting with the agent directly.

Fleet Architecture

Understanding the architecture helps you design efficient fleets:

Fleet Controller (Your Dashboard)
├── Agent Group: Content Team
│   ├── Blog Writer Agent      [Active] → 3 tasks queued
│   ├── Social Media Agent     [Active] → Running task
│   └── Email Newsletter Agent [Idle]
├── Agent Group: Development
│   ├── Frontend Dev Agent     [Active] → Building component
│   ├── Backend API Agent      [Active] → Running tests
│   └── QA Testing Agent       [Idle]   → Awaiting deployment
└── Agent Group: Client Work
    ├── ClientA SEO Agent      [Active] → Site audit
    ├── ClientB Content Agent  [Active] → Writing articles
    └── ClientC Design Agent   [Paused] → Awaiting approval

Each agent group can share context and communicate with each other. The Frontend Dev Agent can pass completed components to the QA Testing Agent automatically, creating a pipeline that mirrors real development workflows.

Advanced Fleet Patterns

The Assembly Line

Set up agents in a sequential pipeline where each agent's output becomes the next agent's input:

1. Research Agent gathers data and creates a brief
2. Writer Agent produces a draft from the brief
3. SEO Agent optimizes the draft for search engines
4. Editor Agent proofreads and polishes the final version
5. Social Agent creates promotional content from the finished piece

This pattern is especially powerful for content agencies. One fleet running this pipeline can produce 20-30 polished, SEO-optimized articles per day — work that would typically require a team of 5-8 people.

The Swarm

Deploy multiple identical agents to parallelize a large task. Need to audit 500 web pages for accessibility? Deploy 10 Accessibility Auditor agents, each processing 50 pages simultaneously. What would take one agent 8 hours completes in under an hour.

The Specialist Team

Create a cross-functional team of specialized agents that collaborate on complex projects. For a website redesign project:

• UX Researcher Agent analyzes current user behavior and competitor sites
• Design System Agent creates a comprehensive design token system
• Frontend Agent builds responsive components using the design tokens
• Performance Agent optimizes assets and code for Core Web Vitals
• Copy Agent writes all page content optimized for conversion

Monitoring and Analytics

The Fleet Dashboard provides real-time visibility:

• Agent Status Map: Visual overview of all agents showing active, idle, paused, and error states
• Token Consumption: Per-agent and per-group token usage with burn rate projections
• Task Queue: View pending, in-progress, and completed tasks across all agents
• Output Gallery: Review all agent outputs from one centralized location
• Performance Metrics: Track average task completion times, quality scores, and efficiency metrics
• Cost Dashboard: Real-time cost tracking with budget alerts and spending forecasts

Fleet API

For programmatic control, the Fleet API lets you manage agents from your own applications:

// Create a new agent via API
const response = await fetch('https://gositeme.com/api/fleet', {
    method: 'POST',
    headers: {
        'Authorization': 'Bearer YOUR_API_KEY',
        'Content-Type': 'application/json'
    },
    body: JSON.stringify({
        name: 'SEO-Auditor-March',
        role: 'seo-specialist',
        tools: ['seo-analyzer', 'keyword-research', 'schema-generator'],
        token_budget: 50000,
        context: { project: 'client-website-redesign' }
    })
});

// Assign a task
await fetch('https://gositeme.com/api/fleet/agents/SEO-Auditor-March/tasks', {
    method: 'POST',
    headers: { 'Authorization': 'Bearer YOUR_API_KEY' },
    body: JSON.stringify({
        task: 'Perform a complete SEO audit of https://example.com',
        output_format: 'pdf-report',
        priority: 'high'
    })
});

Best Practices for Large Fleets

• Start small, scale gradually. Begin with 3-5 agents, optimize their configurations, then replicate successful patterns.
• Use agent groups to organize by client, project, or function. This keeps your dashboard manageable at scale.
• Set token budgets for every agent. Runaway agents without budget limits can consume resources quickly.
• Review output quality weekly. Use the Output Gallery to spot-check agent work and refine configurations.
• Archive completed agents to keep your active fleet focused. Archived agents retain their context for future reuse.
• Use webhooks for task completion notifications instead of polling the dashboard.