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

# Quickstart Guide

> Train and deploy your first Leadblanc AI agent in under 3 minutes.

Get your AI customer assistant up and running in three simple steps.

**1. Create your Workspace & Connect Data**

Log into your [Leadblanc Dashboard](https://app.leadblanc.com).

1. Navigate to **Build > Data Sources**.
2. Enter your public website URL (e.g. `https://yourbrand.com`) and click **Crawl**.
3. Our automated crawler extracts product catalogs, FAQs, return policies, and brand guidelines into clean vector embeddings.

**2. Tune Instructions & Safety Guardrails**

Go to **Build > Instructions**.

* Set the **Agent Personality** (e.g. *Helpful, concise, and focused on consultative product recommendations*).
* Under **Guardrails**, configure prohibited topics, price boundaries, and response behavior for out-of-scope inquiries.
* Test the behavior immediately using **Test in Playground** to simulate live customer conversations.

**3. Deploy to Web Chat or WhatsApp**

When you are satisfied with the responses:

* To embed on your website, copy the one-line snippet from **Channels > Web Chat**.
* To link WhatsApp, connect your Meta Cloud API account under **Channels > WhatsApp**.

***

## Verifying in the Playground

The **Playground** is an isolated simulator where you can test conversational flows, test edge-case queries, and inspect raw reasoning without affecting live customer records or consumption quotas.

```bash theme={null}
# Example test query in Playground
Customer: "Do you offer free shipping on orders above $50 to Canada?"
Agent: "Yes! We offer free standard shipping to Canada on all orders over $50 CAD. Delivery typically takes 3-5 business days."
```

> [!TIP]
>
> You can toggle between **Customer View** and **Debug Inspector** in the Playground to examine which retrieved data sources were referenced in each response.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.