Tasks Overview

Tasks are the fundamental unit of work in Xagent. You describe a goal in natural language and Xagent plans, executes, and delivers the result — no predefined workflow required.

What is a Task?

A task is a plain-language description of an outcome. Xagent interprets your intent and breaks it into executable steps.

"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"

Task Configuration

Before running a task you can choose the models it uses:

  • Main Model — used for all execution by default.
  • Fast Model — used automatically for steps identified as simple.
  • Long Context / Vision — for context compression and image tools.

You can also adjust temperature, max tokens, and other parameters per task. See LLM Models.

Key Features

  • Natural-language input — describe the goal in plain language.
  • File attachments — upload documents, spreadsheets, images, or code for context. See File Upload.
  • Real-time execution — watch each step get planned and executed, view intermediate results, and iterate.

Creating a Task via API

POST /api/chat/task/create
{
  "description": "Summarize the EV battery market",
  "files": []
}

Next Steps