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

# Quickstart

> From request to real execution in three steps.

## Step 1: Install Lukan Agent

```bash theme={null}
curl -fsSL https://get.lukan.ai/install.sh | bash
```

Supported platforms:

* Linux (glibc): x86\_64, arm64
* macOS: x86\_64, arm64 (Apple Silicon)

<Note>
  Alpine/musl and Windows are not supported. Use WSL on Windows.
</Note>

## Step 2: Configure

```bash theme={null}
lukan setup
```

This interactive wizard helps you:

1. Select a model provider (Nebius, Anthropic, Fireworks, GitHub Copilot, OpenAI Codex, z.ai, Ollama Cloud, OpenAI Compatible, Lukan Cloud)
2. Enter your API key
3. Set your timezone

## Step 3: Start the assistant

```bash theme={null}
lukan
```

This opens the interactive CLI chat.

## Step 4: Describe the goal

Use natural language to request concrete outcomes:

```text theme={null}
Open the project, find all TODO comments, and create a short action list.
```

Lukan then plans, connects tools, executes actions, and validates the result.

## Optional: Initialize a project

```bash theme={null}
cd your-project
lukan init
```

This creates a `.lukan/` directory with:

* `config.json` - Project-specific permissions
* `memories/MEMORY.md` - Project context that persists across sessions

## What to do next

<CardGroup cols={2}>
  <Card title="Capabilities" icon="bolt" href="/capabilities/overview">
    Understand how Lukan plans and executes complex workflows.
  </Card>

  <Card title="CLI Reference" icon="book" href="/reference/cli">
    Learn the core commands for daily usage.
  </Card>

  <Card title="Permissions" icon="lock" href="/permissions">
    Configure what Lukan can and cannot do in your project.
  </Card>

  <Card title="Providers" icon="cloud" href="/providers">
    Configure model providers and API keys.
  </Card>
</CardGroup>