Tasks - PraisonAI

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

32 options on Task.__init__ are accepted for Python-SDK parity but not yet acted on in the TypeScript SDK. If you pass one and see a [praisonai] … not yet honoured notice, see the Parity Notices page for the full list.

The behaviour behind 28 of these options is now implemented in the SDK’s Task engine and will become active once the team runner wires it in — see PR #4838. Until then these options still emit the parity notice above.

Quick Start

Simple Usage

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);

With Configuration

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

Option	Description
`description`	What the task should accomplish
`expectedOutput`	Format/structure of result
`agent`	Agent assigned to the task
`context`	Other tasks to use as input
`async`	Run without blocking

API Reference

TaskConfig

Task configuration options

Task

Task class documentation

Best Practices

Be specific in descriptions

“Write 500-word blog post about AI benefits” beats “Write about AI”.

Define expected output

Telling the agent what format you want improves results.

Use context for dependencies

Chain tasks by passing earlier outputs as context.

Workflows

Multi-step workflows

Teams

Multi-agent teams

​Quick Start

​User Interaction Flow

​Configuration Levels

​Task Options

​API Reference

TaskConfig

Task

​Best Practices

​Related

Workflows

Teams

Quick Start

User Interaction Flow

Configuration Levels

Task Options

API Reference

Best Practices

Related