Skip to main content
The code execution action empowers your Agents to write and execute Python code, enabling them to tackle more advanced tasks such as running calculations, analyzing data, and producing visualizations.

​
Capabilities

  • Calculations — Accurately handle large-scale or complex mathematical operations
  • Data Analysis — Process and analyze files uploaded by the user, including CSVs and spreadsheets
  • File Generation — Produce and return files such as reports, updated datasets, or exported content
  • Visualizations — Build graphs and charts that render directly within the Chat UI
To activate it for a specific Agent, simply toggle on the Code Execution action within that Agent’s configuration settings.

​
How it Works

Code execution operates within a secure, sandboxed Python environment that includes:
  • Pre-installed libraries such as numpy, pandas, scipy, and matplotlib
  • Restricted access — network connectivity is disabled and filesystem access is limited for security purposes
  • Resource controls — usage limits and timeouts are in place to prevent uncontrolled execution
The underlying LLM automatically determines when to invoke code execution based on the user’s request — no manual trigger is required.