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

# Editor Overview

The Editor is the primary workspace in Intelleap for creating, configuring, and running prompts.

Each prompt includes three connected workspaces:

* Editor
* Testing
* Evaluations

Work typically begins in the Editor, continues in Testing, and concludes with reviewing results in Evaluations.

<img src="https://mintcdn.com/intelleap/h4CIG-p7Mjr0kT4g/images/screens/Editor/1.png?fit=max&auto=format&n=h4CIG-p7Mjr0kT4g&q=85&s=7e343a9927225f4aafc455590551d3c8" alt="Full Editor Layout" width="3280" height="2202" data-path="images/screens/Editor/1.png" />

## Editor Layout

The Editor consists of two main areas:

<CardGroup cols={2}>
  <Card title="Variables sidebar">
    Manage scenarios and variables associated with the current prompt.
  </Card>

  <Card title="Prompt variants workspace">
    Create, configure, run, and compare one or more prompt variants.
  </Card>
</CardGroup>

The sidebar contains scenarios and variables associated with the current prompt.

The main workspace contains one or more prompt variants.

## Scenarios

Scenarios store different sets of values for prompt variables.

Switching between scenarios changes the values assigned to variables while preserving the variable structure itself.

This makes it possible to test the same prompt variants against multiple input configurations.

New scenarios can be created from the scenario selector located at the top of the sidebar.

## Variables

Variables define the data used throughout prompt variants.

Intelleap supports the following variable types:

<CardGroup cols={3}>
  <Card title="String">
    Plain text values used for inputs, instructions, parameters, and test data.
  </Card>

  <Card title="Chat History">
    Multi-message conversation context with user, assistant, and system roles.
  </Card>

  <Card title="JSON">
    Structured data edited through the built-in JSON editor.
  </Card>
</CardGroup>

Variables can be reordered using drag-and-drop and reused across all prompt variants.

## Prompt Variants

Prompt variants allow different prompt configurations to be tested side by side.

Each variant contains:

* Model settings
* Shared prompt body
* Context Before blocks
* Context After blocks
* Output area

Variants can be created, cloned, updated, or removed independently.

Users without full access can create up to three variants per prompt.

## Running Variants

Prompt variants can be executed individually or together.

<CardGroup cols={2}>
  <Card title="Run">
    Execute one variant and display the result inside that variant.
  </Card>

  <Card title="Run All">
    Execute all visible variants simultaneously from the Editor header.
  </Card>
</CardGroup>

Execution results are displayed directly within each variant.

## Comparing Variants

The **Differences** action allows users to compare the structure of prompt variants.

When multiple variants are available, any two variants can be selected and compared side by side.

Differences are displayed in a dedicated comparison view.

The **Differences** action becomes available only when more than one variant exists.

<img src="https://mintcdn.com/intelleap/h4CIG-p7Mjr0kT4g/images/screens/Editor/2.png?fit=max&auto=format&n=h4CIG-p7Mjr0kT4g&q=85&s=709c227d0f6d7c1f571079f4fd9f0801" alt="Differences Modal" width="3280" height="2202" data-path="images/screens/Editor/2.png" />

## Releases

The **Releases** action creates and manages prompt releases.

A release captures the current state of a prompt and makes it available in Testing and Evaluations.

The Releases dialog provides two options:

* New Release
* Update Release

Creating the first release unlocks access to both the Testing and Evaluations tabs.

## Usage Limits

Users without full access are limited to:

* 3 scenarios per prompt
* 10 variables per prompt
* 3 prompt variants per prompt

<Note>
  These limits apply to users without full access. Full-access workspaces can use higher limits.
</Note>