Skip to main content

​
Files Overview

Files enable agents to work with actual documents and data during task execution. Upload reference materials, data files, or any documents that agents need to analyze, process, or reference.

​
What Are Files?

Files are documents and data that you can upload for agents to use:
  • Analyze - Process documents, extract information
  • Reference - Consult during task execution
  • Transform - Convert between formats
  • Generate - Create new files based on inputs

​
Why Use Files?

Document Analysis

Upload PDFs, Word docs, presentations for analysis

Data Processing

Upload CSV, Excel files for analysis and visualization

Code Review

Upload code files for review and improvement

Context Input

Provide reference material for content generation

​
How Files Work

​
Upload and Process

When you upload a file:
  1. Upload - File is stored in your workspace
  2. Available - File is accessible to agents
  3. Read - Agent reads file when needed
  4. Process - Agent analyzes and processes content
  5. Generate - Agent can create new files as output

​
Workspace

Each task has an isolated workspace:
  • Storage - Files stored in task-specific workspace
  • Access - Only current task can access its workspace
  • Operations - Agents can read, write, and list files
  • Output - Generated files saved to workspace
Files are managed per-task. When you create a task and upload files, they’re stored in that task’s workspace and accessible only during that task execution.

​
Supported File Types

​
Documents

  • PDF - .pdf
  • Word - .doc, .docx
  • Text - .txt
  • Markdown - .md
  • HTML - .html, .htm

​
Spreadsheets & Data

  • Excel - .xlsx, .xls
  • CSV - .csv
  • JSON - .json

​
Presentations

  • PowerPoint - .pptx

​
Images

  • PNG - .png
  • JPG/JPEG - .jpg, .jpeg
  • GIF - .gif
  • SVG - .svg

​
Code

  • Python - .py
  • JavaScript - .js
  • TypeScript - .ts
  • Java - .java
  • And more - Most programming languages

​
Archives

  • ZIP - .zip

​
File Operations

Agents can perform various operations on files:

​
Reading Files

  • Text files - Read content directly
  • Documents - Parse and extract text (PDF, DOCX, etc.)
  • Spreadsheets - Read structured data (CSV, Excel)
  • Images - Use vision tools to analyze

​
Writing Files

  • Create new files - Save output and results
  • Generate reports - Create documents from analysis
  • Export data - Save processed information
  • Code generation - Write code files

​
Listing Files

  • Browse workspace - See available files
  • Check directories - Navigate file structure
  • Find files - Search by pattern

​
Editing Files

  • Replace content - Update existing files
  • Insert content - Add new sections
  • Delete content - Remove parts of files

​
Use Cases

​
Data Analysis

Upload data files for analysis and visualization: 
User: [Uploads sales_data.csv]
User: "Analyze this data and show trends"

Agent:
1. Reads CSV with file tool
2. Uses Python executor for analysis
3. Creates charts and visualizations
4. Saves report with write_file

​
Document Review

Upload documents for summarization and analysis: 
User: [Uploads contract.pdf]
User: "Summarize key terms and risks"

Agent:
1. Uses document_parser to extract text
2. Analyzes content with LLM
3. Identifies key clauses and risks
4. Provides summary

​
Content Generation

Upload reference material for content creation: 
User: [Uploads brand_guidelines.pdf]
User: "Create a presentation following these guidelines"

Agent:
1. Reads brand guidelines
2. Extracts design rules and tone
3. Generates PowerPoint presentation
4. Saves as PPTX file

​
Code Processing

Upload code for review or modification: 
User: [Uploads app.py]
User: "Review this code and suggest improvements"

Agent:
1. Reads Python file
2. Analyzes code structure
3. Identifies issues and improvements
4. Writes updated version

​
File Size Limits

File size limits depend on your configuration:
  • Default limit - Typically 10-100MB per file
  • Plan limits - Higher tiers may allow larger files
  • Recommendation - Compress large files before uploading
For very large files, consider splitting into smaller chunks or using compression.

​
File Security

​
Storage

  • Isolated - Files stored in task-specific workspace
  • Temporary - Files available only during task execution
  • Encrypted - Files encrypted at rest

​
Access Control

  • Per-task - Only current task can access its files
  • User-scoped - Your files are private to your tasks
  • Agent-managed - Agents operate on files within task permissions

​
Privacy

  • No sharing - Files are not shared across tasks
  • Auto-cleanup - Files can be automatically cleaned up after task
  • Secure - Transfers and storage are encrypted

​
Best Practices

​
File Preparation

  • Use descriptive names - Helps agents understand file contents
  • Clean formatting - Well-structured files are easier to process
  • Remove sensitive data - Don’t upload confidential information
  • Check file size - Compress if needed

​
Multiple Files

  • Organize related files - Upload together for context
  • Use clear names - Differentiate similar files
  • Consider order - Upload in logical sequence if needed
  • Provide context - Explain what each file contains

​
After Tasks

  • Download outputs - Save generated files before task closes
  • Review results - Check agent-generated files
  • Archive if needed - Keep important outputs
  • Clean workspace - Delete unnecessary files

​
Next Steps