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

# Data

<Frame>
  <img src="https://mintcdn.com/chatzyai/f4zfTmwUuctQb-Q3/ai-agents/voice-ai-agent/data.gif?s=9c6438c68eccc64c26fa22970f99de83" width="800" height="450" data-path="ai-agents/voice-ai-agent/data.gif" />
</Frame>

Your bot learns from various data sources. The more relevant the data, the better your agent performs.

## **Supported Data Types:**

* **Text:** Paste plain text directly into the platform.
* **Files:** Upload formats like `.pdf`, `.docx`, `.txt`, `.csv` (max 20MB).
* **Website URLs:** Add a URL and let Chatzy extract useful content.
* **FAQs:** Manually input question-answer pairs for precise responses.
* **Google Drive:** Link a public folder with view-only access.

***

### Website Data Options

When adding **Website URLs** as a data source, Chatzy provides multiple ways to fetch content:

* **Crawl**\
  Automatically scans all pages linked from the given URL (excluding file downloads).\
  Example: entering `https://example.com` will crawl all linked pages under that domain.

* **Sitemap**\
  Provide an XML sitemap URL to fetch structured content from your website.

* **Link**\
  Add a single webpage URL to pull specific content.

* **CSV File**\
  Upload a CSV that contains multiple URLs for bulk training.

***

**Apify Crawler (Optional)**

You can toggle **Use Apify Crawler** for a more advanced and accurate crawling option.

* It consumes **2 credits per webpage crawled**.
* Designed to extract cleaner, structured data optimized for LLM training.
* Learn more: [Apify Content Crawler](https://apify.com/apify/website-content-crawler).

***

> The Website tab gives you flexibility — crawl entire sites, upload sitemaps, add single links, or bulk import via CSV. For maximum accuracy, enable **Apify Crawler**.

***

### FAQ CSV Q\&A Import

You can also bulk import FAQs by uploading a **CSV file**:

* The **first column** will be treated as the *question*.
* The **second column** will be treated as the *answer*.
* The file **must not contain headers**.

This is useful for quickly loading a large set of predefined Q\&A pairs into your agent.

***

### Google Drive Import

You can connect your Google Drive to train your agent with files:

* Enter the **view-only link** of a Google Drive folder containing your **Sheets or Documents**.
* The system will process all files inside the folder.
* Ensure the folder is **publicly accessible** (view-only).

This allows seamless training from existing documents stored in Google Drive.

***

**🚀 Training Steps:**

1. Upload or connect your data.
2. Click **Train** in the top-right corner.
3. Wait for the progress bar to complete.
4. Once training finishes, your bot is ready to use the data in live chats.

***

### **Source Type Summary**

| Source Type  | Example Formats             | Max Size |
| ------------ | --------------------------- | -------- |
| Text         | Manual paste                | —        |
| File Upload  | PDF, DOCX, TXT, CSV         | 20 MB    |
| Website URL  | Link to be crawled for data | —        |
| FAQ          | Q\&A pairs for knowledge    | —        |
| Google Drive | Folder link with docs       | —        |