> ## Documentation Index
> Fetch the complete documentation index at: https://docs.hunch.tools/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Introduction

> Execute custom code within your AI workflows

## Overview

The Code Runner is a powerful block type in Hunch that allows users to execute custom code as part of their AI workflows. It bridges the gap between AI-generated content and programmatic actions, enabling more complex and dynamic interactions within your canvas.

## What is a Code Runner?

A Code Runner block is a specialized component that can execute arbitrary Python or Bash code within a secure sandbox environment. It allows users to:

* Process data from other blocks
* Perform calculations or data transformations
* Interact with external APIs or services
* Generate dynamic content or files

Code Runners expand the capabilities of your Hunch canvas beyond what's possible with standard AI blocks alone, allowing for true programmatic control and customization.

## Key Features

* **Language Support**: Execute Python or Bash code directly within your workflow.
* **Secure Execution**: Code runs in an isolated sandbox environment for safety.
* **Input Integration**: Easily access data from connected input blocks.
* **Output Flexibility**: Generate text, files, or other data types as output.
* **Package Management**: Install and use Python packages as needed.

## The Power of Code in AI Workflows

By incorporating Code Runners, Hunch enables a seamless blend of AI capabilities and traditional programming. This combination allows for more sophisticated, flexible, and powerful workflows that can adapt to a wide range of use cases and requirements.

Whether you're a seasoned developer or just getting started with coding, Code Runners provide a flexible and powerful tool to enhance your AI-driven projects in Hunch.