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

# Agent Diff

> Interactive environments for evaluating AI agents & RL training on replicas of 3rd party APIs.

<Tip>
  **Try it now** — Copy one of the Jupyter notebooks:

  * [Diffs Demo](https://colab.research.google.com/drive/1-GZlU-wgAk0W2v8cpQuPP_UbCNJLmVk5?usp=sharing) — Capture agent actions as database diffs
  * [Evaluations Demo](https://colab.research.google.com/drive/1cfeMQ2R_JpGRdagT0U-D8cngpsHJmJON?usp=sharing) — Run assertions against agent diffs
  * [Linear Bench](https://colab.research.google.com/drive/1Hext-WWDsm9BxsOrASYoMjgu1N_lN0Fz) — Run the 40-task benchmark from HuggingFace
</Tip>

## Core Concepts

<CardGroup cols={2}>
  <Card title="Templates & Environments" icon="copy" href="/core-concepts/environments">
    Templates are populated database schemas. Environments are isolated, ephemeral copies of templates where your agents operate. **Each environment gets its own base API URL** that you can proxy to your agents.
  </Card>

  <Card title="Runs & Diffs" icon="code-compare" href="/core-concepts/diffs">
    A run represents a single test session within an environment. **Starting a run takes a snapshot, and ending it takes another one and returns a diff** - computed difference between the before and after states of an environment.
  </Card>

  <Card title="Evaluations" icon="check" href="/core-concepts/evaluations">
    Evaluations let you verify that your AI agent performed the expected actions. You can create your own test suites that will **compare expected state change to diff results**, or use our example ones for Linear and Slack.
  </Card>

  <Card title="Tests & Assertions" icon="check" href="/core-concepts/creating-suites">
    Define **test suites** with expected outcomes using our assertion DSL. Each test specifies a prompt, environment template, and assertions that verify the agent made the correct database changes (inserts, updates, deletes).
  </Card>
</CardGroup>

## Supported APIs

<CardGroup cols={2}>
  <Card title="Slack" icon="slack" href="/services/slack/overview">
    Web API coverage for conversations, chat, reactions, users, and more
  </Card>

  <Card title="Linear" icon="square-kanban" href="/services/linear/overview">
    GraphQL API for issues, teams, projects, comments, and workflow states
  </Card>
</CardGroup>

<Note>
  More APIs coming soon. [Request an integration →](https://github.com/hubertpysklo/agent-diff/issues)
</Note>

## Next Steps

<CardGroup cols={2}>
  <Card title="Quickstart" icon="rocket" href="/quickstart">
    Get up and running in 5 minutes
  </Card>

  <Card title="Example Benchmarks" icon="chart-line" href="/test-suites/benchmarks">
    See built-in evaluation suites
  </Card>

  <Card title="Python SDK" icon="python" href="/sdks/python/installation">
    Full SDK reference for Python
  </Card>

  <Card title="TypeScript SDK" icon="js" href="/sdks/typescript/installation">
    Full SDK reference for TypeScript/
  </Card>
</CardGroup>