Tasks Overview
Tasks are the fundamental unit of work in Xagent. Simply describe what you want to accomplish, and Xagent will plan, execute, and deliver results automatically.What is a Task?
A task is a natural language description of a goal you want to achieve. Unlike traditional automation systems that require predefined workflows, Xagent interprets your intent and breaks it down into executable steps.Task Examples
Task Configuration
Before running a task, you can configure the model settings.
Model Selection
Choose the appropriate model for your task: Main Model - Used for all task execution by default- Required for all tasks
- Handles planning, reasoning, and execution
- Used unless fast model is configured
- Automatically used for steps identified as simple during planning
- Reduces cost and improves speed for routine operations
- If not configured, main model handles all steps
- Long Context Models - For compressing large context windows
- Vision Models - Required to use image analysis tools
- Select from configured providers
- Adjust temperature for creativity
- Set max tokens for output length
- Configure other model parameters

Key Features
Natural Language Input
Describe your goal in plain language - no coding or workflow design required. Xagent understands context and intent.File Attachments
Upload files to provide context for your tasks:- Documents (PDF, DOCX, TXT)
- Spreadsheets (CSV, XLSX)
- Images (PNG, JPG)
- Code files
- And more
Real-time Execution
Watch your task execute in real-time:- See each step as it’s planned
- Monitor execution progress
- View intermediate results
- Iterate on the output
Streaming Responses
Get results as they’re generated, not just at the end. Perfect for long-running tasks where you want to see progress.How It Works
Task Interface
The task interface provides:- Chat Input: Natural language input area
- Model Selector: Choose models for this task
- File Upload: Attach files for context
- Execution View: Watch your task progress in vertical step flow
- Results Panel: Review outputs and artifacts
Next Steps
- Learn about task execution - See how tasks execute step-by-step
- Building agents - Convert tasks to reusable agents
- LLM Models - Set up and configure language models