Skip to main content
Tasks define specific work for agents to complete with clear objectives.

​
Quick Start

1

Simple Task

import { Agent, Task } from 'praisonai';

const agent = new Agent({
  instructions: 'You research and summarize topics'
});

const task = new Task({
  description: 'Research AI trends in 2024',
  agent: agent
});

const result = await task.execute();
console.log(result.output);
2

With Expected Output

const task = new Task({
  description: 'Analyze sales data',
  expectedOutput: 'A summary with top 3 insights',
  agent: analyst
});

​
User Interaction Flow

​
Configuration Levels

// Level 1: String - Simple description
const task = new Task({
  description: 'Write a blog post about AI',
  agent: writer
});

// Level 2: Dict - With options
const task = new Task({
  description: 'Write a blog post',
  expectedOutput: '500 words, engaging tone',
  agent: writer
});

// Level 3: Instance - Full control
const task = new Task({
  description: 'Write a blog post',
  expectedOutput: '500 words',
  agent: writer,
  context: [researchTask],  // Use output from other tasks
  callback: (result) => console.log('Done!')
});

​
Task Options

OptionDescription
descriptionWhat the task should accomplish
expectedOutputFormat/structure of result
agentAgent assigned to the task
contextOther tasks to use as input
asyncRun without blocking

​
API Reference

TaskConfig

Task configuration options

Task

Task class documentation

​
Best Practices

“Write 500-word blog post about AI benefits” beats “Write about AI”.
Telling the agent what format you want improves results.
Chain tasks by passing earlier outputs as context.

Workflows

Multi-step workflows

Teams

Multi-agent teams