# Quickstart
Source: https://docs.treeship.dev/guides/quickstart

> From zero to a verifiable session receipt in under 90 seconds.

<Callout type="info">
  No account, email, or API key required. Everything runs locally.
</Callout>

## One-liner setup

```bash
curl -fsSL treeship.dev/setup | sh
```

This installs the CLI, initializes your ship, and instruments any detected agent frameworks (Claude Code, Cursor, Hermes, OpenClaw).

## Or step by step

<Steps>
  <Step>
    ### Install

    ```bash
    curl -fsSL treeship.dev/install | sh
    ```
  </Step>

  <Step>
    ### Initialize

    ```bash
    treeship init
    ```

    ```
    ✓ Treeship initialized
      Ship ID:  ship_xxx
      Key ID:   key_xxx
    ```
  </Step>

  <Step>
    ### Instrument your agents (optional)

    ```bash
    treeship add
    ```

    Auto-detects Claude Code, Cursor, Hermes, OpenClaw and configures Treeship integration.
  </Step>

  <Step>
    ### Start a session

    ```bash
    treeship session start --name "my first task"
    ```
  </Step>

  <Step>
    ### Wrap commands

    ```bash
    treeship wrap -- npm test
    treeship wrap -- git commit -m "fix bug"
    ```

    Every wrapped command gets a signed artifact with exit code, duration, and output digest.
  </Step>

  <Step>
    ### Close and create your receipt

    ```bash
    treeship session close \
      --summary "Fixed the auth bug and added a regression test"
    ```

    This creates a `.treeship` package with a self-verifying `preview.html` that opens in your browser.
  </Step>

  <Step>
    ### Verify locally

    ```bash
    treeship package verify .treeship/sessions/ssn_*.treeship
    ```

    Should show all PASS. This runs offline with zero network calls.
  </Step>

  <Step>
    ### Share (optional)

    ```bash
    treeship hub attach --endpoint https://api.treeship.dev
    treeship session report
    ```

    Prints a permanent public URL like `https://treeship.dev/receipt/ssn_xxx`. No auth needed to view it.
  </Step>
</Steps>

## Interactive quickstart

For a fully guided experience:

```bash
treeship quickstart
```

Walks through init, wrapping a command, and creating a receipt interactively.

## Environment variables for richer receipts

Set these before running commands to capture model and token data:

```bash

```

<Callout type="info">
  Cost is not captured in receipts. Pricing depends on provider, tier, and timing. Receipts store verifiable usage (model + tokens). Cost is calculated by consumer tools (Witness dashboards, billing) from their pricing config.
</Callout>

## Live monitoring

While your agent works, watch the session in real time:

```bash
treeship session status --watch
```

## What's next

* [Session Receipts](/docs/concepts/session-receipts) -- understand the receipt format
* [Claude Code integration](/docs/integrations/claude-code) -- setup for Claude Code
* [CLI reference](/docs/cli/overview) -- all commands