How to Get Started with Emdash: Run Multiple Coding Agents in Parallel

Getting Started with Emdash: A Hands-On Tutorial

In this tutorial, we'll walk through setting up Emdash and running your first multi-agent coding session. By the end, you'll have a fully working environment where multiple AI agents collaborate on your codebase.

Step 1: Deploy with Docker

The fastest way to get started is with the official Docker image:

docker run -d \
  --name emdash \
  -p 8080:8080 \
  -v ./data/emdash:/data \
  --restart unless-stopped \
  jstgnkl/emdash:latest

Or use Docker Compose for easier management:

services:
  emdash:
    image: jstgnkl/emdash:latest
    restart: unless-stopped
    ports:
      - 8080:8080
    volumes:
      - ./data/emdash:/data

Step 2: Configure AI Providers

Emdash supports multiple AI providers out of the box. You'll need API keys for the providers you want to use. Add them in the Emdash settings interface:

Step 3: Run Your First Multi-Agent Session

Once configured, create a new project and assign agents to tasks:

1. Create a project — Define your codebase directory and goals
2. Spawn agents — Launch 2-3 agents, each with a different provider and role
3. Assign tasks — Give Agent 1 a feature implementation, Agent 2 test writing, Agent 3 code review
4. Monitor in real-time — Watch all agents work simultaneously in the Emdash interface

Real-World Use Cases

Tips for Best Results

• Use different providers for different tasks — Claude excels at reasoning, GPT-4o at creative tasks, Gemini at long-context analysis
• Set clear boundaries between agent tasks to avoid conflicts
• Use the terminal integration for real-time feedback loops
• Start with 2-3 agents and scale up as you learn the workflow