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

# Scale Gymnasium

> Standardized environments for training and evaluating LLM agents

Scale Gymnasium provides a standardized collection of digital environments for agent training and evaluation. Whether you're building tool-use agents, or GUI agents for browser-use or desktop-use, Gymnasium offers the environments, data, and verification systems you need.

## Environments

<CardGroup cols={3}>
  <Card title="Website Environments" icon="globe" href="/environments/website/overview">
    **4 synthetic web apps**

    Calendr · Cloudfile · Shopora · Pandora's Inbox

    Full-stack websites with databases, logging, and built-in verification
  </Card>

  <Card title="Desktop Environments" icon="desktop" href="/environments/desktop/overview">
    **3 VM platforms**

    Windows · Ubuntu · macOS

    Isolated virtual machines for computer-use agent testing
  </Card>

  <Card title="MCP Environment" icon="wrench" href="/environments/mcp/overview">
    **45+ servers · 300+ tools**

    Calendar · CRM · Email · Shopping · and more

    Tool-based agent interactions via Model Context Protocol
  </Card>
</CardGroup>

## Two Ways to Use Scale Gymnasium

| Method               | Agent Loop                                       | Best For                                          |
| -------------------- | ------------------------------------------------ | ------------------------------------------------- |
| **Gymnasium Web UI** | Built-in — Scale provides an agent loop executor | Prototyping, testing, exploration                 |
| **Docker Images**    | Bring your own — you implement the agent loop    | Large-scale evaluation, CI/CD, training pipelines |

<CardGroup cols={2}>
  <Card title="Web UI Guide" icon="play" href="/getting-started/web-ui-guide">
    Try environments instantly through the Gymnasium interface
  </Card>

  <Card title="Docker Quick Start" icon="docker" href="/getting-started/quickstart-docker">
    Run environments locally with your own agent loop
  </Card>
</CardGroup>

## Key Capabilities

<CardGroup cols={3}>
  <Card title="Session Isolation" icon="shield">
    Complete data isolation between test sessions ensures no cross-contamination
  </Card>

  <Card title="Comprehensive Logging" icon="list">
    Automatic capture of all agent interactions for analysis and debugging
  </Card>

  <Card title="Built-in Verification" icon="check">
    State checks, log validation, and rubric-based evaluation framework
  </Card>
</CardGroup>

## What's Included

| Resource           | Description                                                          |
| ------------------ | -------------------------------------------------------------------- |
| **Environments**   | 4 website apps + 3 desktop VMs + 45 MCP servers                      |
| **Data Packs**     | Pre-configured datasets to hydrate environments with realistic state |
| **Verifiers**      | State checks, log checks, and rubric-based evaluation                |
| **Sample RL Data** | Pre-built tasks with prompts and verification criteria               |
| **APIs**           | HTTP endpoints for programmatic environment control                  |

## Next Steps

<CardGroup cols={2}>
  <Card title="Key Concepts" icon="book" href="/getting-started/key-concepts">
    Understand Environment, Task, Artifact, and Verifier terminology
  </Card>

  <Card title="Choose Your Path" icon="route" href="/getting-started/choose-your-path">
    Decide between Web UI and Docker based on your needs
  </Card>
</CardGroup>