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

# Quickstart

> Go from zero to a running, verifiable AI agent in under 2 minutes.

import { Steps, CodeGroup, Tip, Warning, Card, CardGroup } from 'mintlify/components'

# Launch Your First Agent

Pick the path that fits you. Both end in the same place: a live AI agent with an on-chain passport, cryptographic receipts, and Solana-anchored proof of every inference.

<CardGroup cols={2}>
  <Card title="Telegram (fastest)" icon="paper-plane" href="#path-1-telegram">
    No terminal. No API key. Tap a few buttons and your agent is live.
  </Card>

  <Card title="CLI (developers)" icon="terminal" href="#path-2-cli">
    One install, one login, one command. Full control from your terminal.
  </Card>
</CardGroup>

***

## Path 1: Telegram

The fastest way to launch. You only need a Telegram account.

<Steps>
  <Step title="Open the Lucid bot">
    Go to [t.me/mylclaw\_bot](https://t.me/mylclaw_bot) in Telegram and tap **Start**.
  </Step>

  <Step title="Start the launch wizard">
    Send `/launch`. The bot walks you through everything.
  </Step>

  <Step title="Pick an agent">
    Tap **OpenClaw** from the catalog. It is a general-purpose conversational agent ready to go out of the box.
  </Step>

  <Step title="Choose your model provider">
    Select **Lucid Gateway (no key needed)**. Lucid routes your calls across 100+ models with zero configuration on your part.
  </Step>

  <Step title="Choose a deploy target">
    Tap **Lucid Cloud (one-click)**. Your agent will be deployed to managed infrastructure in seconds.
  </Step>

  <Step title="Add skills (optional)">
    The bot offers optional skills (web search, code execution, etc.). Skip them for now or select a few — you can change this later.
  </Step>

  <Step title="Launch">
    Tap **Launch** and wait about 30 seconds. The bot will confirm once your agent is live.
  </Step>

  <Step title="Talk to your agent">
    Send any message in the chat. Your agent responds — and every response is backed by a verifiable receipt.
  </Step>
</Steps>

<Tip>
  That is it. You just deployed a verifiable AI agent without writing a single line of code.
</Tip>

***

## Path 2: CLI

For developers who want terminal-level control.

<Steps>
  <Step title="Install the Lucid CLI">
    <CodeGroup>
      ```bash npm theme={null}
      npm i -g @lucid/cli
      ```

      ```bash yarn theme={null}
      yarn global add @lucid/cli
      ```

      ```bash pnpm theme={null}
      pnpm add -g @lucid/cli
      ```
    </CodeGroup>
  </Step>

  <Step title="Authenticate">
    ```bash theme={null}
    lucid login
    ```

    This opens your browser for a one-time auth flow. Your session is stored locally.
  </Step>

  <Step title="Launch an agent">
    ```bash theme={null}
    lucid launch --agent openclaw --target railway
    ```

    The CLI pulls the OpenClaw image, provisions a Railway service, registers an on-chain passport, and deploys — all in one command.

    You will see output like this:

    ```
    ✔ Passport created: LCDp...7xKm
    ✔ Image pushed to registry
    ✔ Railway service provisioned
    ✔ Agent live at https://openclaw-a3f9.up.railway.app

    Your agent is running. Every inference is receipted and anchored to Solana.
    ```
  </Step>

  <Step title="Verify it is running">
    ```bash theme={null}
    lucid deploy status LCDp...7xKm
    ```

    ```
    Status:     RUNNING
    Target:     railway
    Passport:   LCDp...7xKm
    Uptime:     2m 14s
    Receipts:   0
    ```
  </Step>

  <Step title="Send your first message">
    ```bash theme={null}
    curl -X POST https://openclaw-a3f9.up.railway.app/v1/chat/completions \
      -H "Content-Type: application/json" \
      -d '{
        "model": "gpt-4o",
        "messages": [{"role": "user", "content": "Hello from Lucid!"}]
      }'
    ```

    The response includes a `x-lucid-receipt-id` header — that is your cryptographic proof.
  </Step>
</Steps>

***

## What Just Happened?

Whether you used Telegram or the CLI, behind the scenes Lucid did all of this:

<Steps>
  <Step title="Passport minted">
    Your agent received an **on-chain passport** — a unique identity registered on Solana. This is not a database entry; it is a verifiable on-chain record that proves your agent exists and who owns it.
  </Step>

  <Step title="Agent deployed">
    The agent container was built, pushed to a registry, and deployed to your chosen target (Lucid Cloud or Railway). It is running right now, ready to handle requests.
  </Step>

  <Step title="Inference receipted">
    Every message your agent processes generates a **cryptographic receipt** — a SHA-256 hash of the canonicalized request and response, stored in a Merkle Mountain Range.
  </Step>

  <Step title="Receipts anchored to Solana">
    Receipts are batched into **epochs**. When an epoch closes (100+ receipts or 1 hour), the MMR root is committed on-chain. Anyone can verify any receipt against this anchor.
  </Step>

  <Step title="Reputation begins">
    Your agent is now building reputation. Every successful inference, every verified receipt, every on-chain anchor contributes to a reputation score that other agents, users, and protocols can query.
  </Step>
</Steps>

```
You / Telegram / curl
        │
        ▼
┌──────────────┐     ┌──────────────┐     ┌──────────────┐
│  Your Agent  │────▶│  Lucid       │────▶│  LLM Model   │
│  (Passport)  │     │  Gateway     │     │  (100+)      │
└──────────────┘     └──────┬───────┘     └──────────────┘
                            │
                     ┌──────▼───────┐
                     │   Receipt    │  SHA-256 + MMR proof
                     └──────┬───────┘
                            │
                     ┌──────▼───────┐
                     │   Solana     │  On-chain epoch anchor
                     └──────────────┘
```

***

## Next Steps

<CardGroup cols={2}>
  <Card title="Add Skills" icon="puzzle-piece" href="/deploy/setup-wizard">
    Give your agent web search, code execution, and custom tool access.
  </Card>

  <Card title="Understand Passports" icon="id-card" href="/concepts/passports">
    Learn how on-chain identity works for agents, models, compute, tools, and datasets.
  </Card>

  <Card title="Explore Receipts" icon="receipt" href="/concepts/receipts">
    Dive into cryptographic receipts, MMR proofs, and epoch anchoring.
  </Card>

  <Card title="Deploy from Source" icon="code" href="/deploy/from-source">
    Build a fully custom agent with your own code and deploy it through Lucid.
  </Card>
</CardGroup>

<Tip>
  Need help? Message the bot `/help` in Telegram or run `lucid --help` in your terminal. You can also reach us at [kevin.wayne@raijinlabs.io](mailto:kevin.wayne@raijinlabs.io).
</Tip>