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

# Prompt Studio

> Centralized prompt management with version control, testing, and API access.

**Prompt Studio** is Traceport's dedicated environment for prompt engineering. It provides an IDE-like experience for creating, testing, versioning, and publishing prompts — all from a single interface.

<Frame>
  <video autoPlay loop muted playsInline src="https://mintcdn.com/traceport/ZXBNThc7kv_3RvjA/images/prompts.mp4?fit=max&auto=format&n=ZXBNThc7kv_3RvjA&q=85&s=7af9803568045bddae75206bb48705e5" className="w-full rounded-lg" data-path="images/prompts.mp4" />
</Frame>

## Why Use Prompt Studio?

<CardGroup cols={2}>
  <Card title="Decouple Prompts from Code" icon="code-branch">
    Update prompts without redeploying your application. No more hardcoded system instructions.
  </Card>

  <Card title="Version Control" icon="clock-rotate-left">
    Track every change with full version history. Roll back to any previous version instantly.
  </Card>

  <Card title="Multi‑Model Testing" icon="flask-vial">
    Test the same prompt across multiple models simultaneously. Compare quality, latency, and cost.
  </Card>

  <Card title="Automated Evaluation" icon="chart-simple">
    Define evaluation rules that automatically score prompt outputs for quality, relevance, and safety.
  </Card>

  <Card title="Team Collaboration" icon="users">
    Comments, history tracking, and shared prompts enable team-wide prompt engineering.
  </Card>

  <Card title="API Access" icon="code">
    Publish prompts and call them via the REST API using a unique slug identifier.
  </Card>
</CardGroup>

## Prompt Studio Sidebar

The Prompt Studio has its own sidebar with these sections:

| Section      | Description                                   |
| ------------ | --------------------------------------------- |
| **Editor**   | The main prompt authoring workspace           |
| **Comments** | Team discussions and feedback on the prompt   |
| **History**  | Change log showing who modified what and when |
| **Versions** | Draft and published version management        |
| **Optimize** | AI-powered prompt optimization suggestions    |
| **Datasets** | Manage test datasets for batch evaluation     |

## Creating a Prompt

<Steps>
  <Step title="Create">
    Navigate to **Prompts** in the sidebar and click **Create New Prompt**. Give it a name and the system generates a unique slug.
  </Step>

  <Step title="Author">
    Build your prompt using structured [message blocks](/prompt-studio/editor) — System, User, and Assistant messages.
  </Step>

  <Step title="Test">
    [Run your prompt](/prompt-studio/testing) against one or multiple models to validate the output.
  </Step>

  <Step title="Evaluate">
    Define [evaluation rules](/prompt-studio/evaluations) and datasets to systematically grade prompt quality.
  </Step>

  <Step title="Publish">
    Publish a version to make it available via the [Run Prompt API](/api-reference/endpoint/run-prompt).
  </Step>
</Steps>