Quickstart

Quick Start Guide

Get up and running with Dispersl in minutes. This guide will walk you through setting up your first multi-agent AI development team.

Prerequisites

Before you begin, ensure you have:

  • Node.js 18+ installed on your system
  • npm or pnpm package manager
  • A Dispersl account (sign up at dispersl.com (opens in a new tab))
  • Your API key from the Dispersl dashboard

Installation

Option 1: CLI Installation (Recommended)

Install the Dispersl CLI globally:

npm install -g dispersl
# or
pnpm add -g dispersl

Option 2: Project-Specific Installation

Add Dispersl to your existing project:

npm install dispersl
# or
pnpm add dispersl

Option 3: MCP Integration

For AI development tools that support MCP (Model Context Protocol):

npm install dispersl-mcp

Authentication

Get Your API Key

  1. Visit the Dispersl Dashboard (opens in a new tab)
  2. Navigate to Settings → API Keys
  3. Click Generate New Key
  4. Copy your API key (keep it secure!)

Configure Authentication

Set your API key as an environment variable:

# Add to your shell profile (.bashrc, .zshrc, etc.)
export DISPERSL_API_KEY="your_api_key_here"
 
# Or create a .env file in your project
echo "DISPERSL_API_KEY=your_api_key_here" > .env

Alternatively, authenticate via CLI:

dispersl auth login
# Follow the prompts to enter your credentials

Your First Project

Initialize a New Project

Create a new project with Dispersl:

# Create and navigate to your project directory
mkdir my-dispersl-project
cd my-dispersl-project
 
# Initialize Dispersl in your project
dispersl init
 
# Follow the interactive setup

The initialization will create:

  • dispersl.config.js - Configuration file
  • .dispersl/ - Local agent data and cache
  • README.md - Project documentation

Configure Your Project

Edit dispersl.config.js to customize your setup:

module.exports = {
  // Project settings
  name: "my-dispersl-project",
  description: "My first multi-agent project",
  
  // Agent configuration
  agents: {
    coding: {
      model: "anthropic/claude-3-sonnet",
      temperature: 0.1,
      maxTokens: 4000
    },
    testing: {
      model: "openai/gpt-4",
      temperature: 0.2,
      maxTokens: 3000
    },
    docs: {
      model: "google/gemini-pro",
      temperature: 0.3,
      maxTokens: 2000
    }
  },
  
  // Integration settings
  integrations: {
    github: {
      enabled: true,
      repository: "your-username/your-repo"
    },
    deployment: {
      platform: "vercel", // or "netlify", "aws", etc.
      autoDeployPR: true
    }
  }
}

Basic Usage

Start Your AI Development Team

Launch your multi-agent team:

dispersl start

This starts the Dispersl daemon and makes your agents available for tasks.

Your First Task

Create a simple web application:

dispersl task create "Build a React todo app with authentication"

Or use the interactive mode:

dispersl chat
> Build a React todo app with authentication and a REST API backend

Monitor Progress

Watch your agents work:

# View active tasks
dispersl tasks list
 
# Monitor a specific task
dispersl task watch <task-id>
 
# View agent logs
dispersl logs --agent coding

MCP Integration Setup

For Cursor IDE

  1. Create or edit ~/.cursor/mcp.json:
{
  "mcpServers": {
    "dispersl": {
      "url": "https://dispersl.com/mcp/YOUR_DISPERSL_MCP_KEY"
    }
  }
}
  1. Get your MCP key from the dashboard:
dispersl mcp key
  1. Restart Cursor and verify the connection:
Ask Cursor: "List my Dispersl integrations"

For VS Code

  1. Install the Dispersl extension from the marketplace
  2. Configure your API key in VS Code settings
  3. Use the command palette: Dispersl: Connect to Team

For Other AI Tools

Check our MCP Integration Guide for setup instructions for:

  • Claude (Anthropic)
  • Windsurf
  • Cline
  • Zed
  • And more...

Example Workflows

Full-Stack Development

# Create a complete web application
dispersl task create "Build a full-stack e-commerce site with:
- React frontend with TypeScript
- Node.js/Express backend
- PostgreSQL database
- Stripe payment integration
- User authentication
- Admin dashboard
- Deployment to AWS"

API Development

# Generate a REST API
dispersl task create "Create a RESTful API for a blog platform with:
- User management and authentication
- CRUD operations for posts and comments
- File upload for images
- Rate limiting and security
- OpenAPI documentation
- Docker containerization"

Testing & Quality Assurance

# Add comprehensive testing
dispersl task create "Add testing to my existing project:
- Unit tests with Jest
- Integration tests for API endpoints
- E2E tests with Playwright
- Code coverage reporting
- CI/CD pipeline setup"

Common Commands

Project Management

# Initialize new project
dispersl init [project-name]
 
# Start agent team
dispersl start
 
# Stop agent team
dispersl stop
 
# Project status
dispersl status

Task Management

# Create new task
dispersl task create "description"
 
# List all tasks
dispersl tasks list
 
# View task details
dispersl task show <task-id>
 
# Cancel task
dispersl task cancel <task-id>

Agent Management

# List available agents
dispersl agents list
 
# View agent status
dispersl agent status <agent-name>
 
# Configure agent
dispersl agent config <agent-name>
 
# View agent logs
dispersl logs --agent <agent-name>

Integration Commands

# Connect GitHub repository
dispersl github connect <repo-url>
 
# Deploy to platform
dispersl deploy <platform>
 
# Manage MCP connections
dispersl mcp list
dispersl mcp connect <tool-name>

Configuration Options

Agent Models

Dispersl supports multiple LLM providers:

// In dispersl.config.js
agents: {
  coding: {
    model: "anthropic/claude-3-sonnet",    // Anthropic
    // model: "openai/gpt-4",              // OpenAI
    // model: "google/gemini-pro",         // Google
    // model: "deepseek/deepseek-r1",      // DeepSeek
  }
}

Performance Tuning

// Optimize for your use case
performance: {
  maxConcurrentTasks: 3,
  agentTimeout: 300, // seconds
  retryAttempts: 2,
  cacheEnabled: true
}

Security Settings

security: {
  encryptLocalData: true,
  requireMFA: false,
  allowedDomains: ["your-domain.com"],
  rateLimiting: {
    enabled: true,
    requestsPerMinute: 100
  }
}

Troubleshooting

Common Issues

Authentication Failed

# Verify your API key
dispersl auth status
 
# Re-authenticate
dispersl auth login

Agent Not Responding

# Check agent status
dispersl agent status coding
 
# Restart specific agent
dispersl agent restart coding
 
# View detailed logs
dispersl logs --agent coding --verbose

Task Stuck or Failed

# View task details
dispersl task show <task-id>
 
# Retry failed task
dispersl task retry <task-id>
 
# Cancel and recreate
dispersl task cancel <task-id>
dispersl task create "updated description"

Getting Help

Debug Mode

Enable verbose logging for troubleshooting:

# Enable debug mode
export DISPERSL_DEBUG=true
 
# Or use the debug flag
dispersl --debug task create "description"

Next Steps

Now that you have Dispersl set up, explore these advanced features:

  1. Multi-Agent Coordination - Learn how agents collaborate
  2. API Integration - Integrate Dispersl into your existing tools
  3. MCP Setup - Connect with AI development environments
  4. Advanced Configuration - Customize agents and workflows
  5. Deployment Guide - Production deployment strategies

Example Projects

Check out these example projects to see Dispersl in action:

Ready to build something amazing? Start with dispersl init and let your AI development team handle the rest!


Need help? Join our Discord community (opens in a new tab) or check out the full documentation.