Skip to main content

​
Quick Start

Get Xagent running and create your first task in 5 minutes.

​
Step 1: Install Xagent

If you haven’t installed Xagent yet, follow the Installation Guide.
For Docker (fastest): 
git clone https://github.com/xorbitsai/xagent.git
cd xagent
cp example.env .env
docker compose up -d
Open http://localhost:80 in your browser.

​
Step 2: Login

Default credentials:
  • Username: administrator
  • Password: administrator

​
Step 3: Configure LLM

Before creating tasks, you need to configure an LLM provider.
  1. Click Models in the left sidebar
  2. Add your API key (OpenAI, Anthropic, or others)
  3. Select a default model
Xagent needs an LLM to plan and execute tasks. See LLM Models for details.

​
Step 4: Create Your First Task

  1. Go to the Task page
  2. Type your request in the chat box
Try one of these examples: 
"Research the latest AI developments and write a summary"
"Create a 10-slide presentation about renewable energy"
"Design a promotional poster for a coffee shop"
  1. Press Enter or click Send
  2. Watch Xagent execute the task in real-time

​
Step 5: View Execution

Xagent will:
  1. Plan the task automatically
  2. Show the DAG (workflow) visualization
  3. Execute each step
  4. Display results
You can:
  • Click nodes in the DAG to see details
  • Send follow-up messages
  • Upload additional files
  • Review execution history

​
What’s Next?

Learn More About Tasks

Understand how tasks work and how to get the most out of them

Build an Agent

Convert your tasks into reusable agents

Explore Templates

Use pre-built templates for common workflows

Add Knowledge Base

Give your agents domain-specific knowledge

​
Troubleshooting

​
Task not starting

Check that:
  • LLM API key is configured correctly
  • Model is selected in Models settings
  • Your request is clear and specific

​
Execution stuck

  • Wait longer (some tasks take time)
  • Check the DAG for stuck steps
  • Click cancel and try again

​
API errors

  • Verify your API key is valid
  • Check API rate limits
  • Try a different model

​
Next Steps