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

# How to add data sources to the Knowledge Base

> Train your AI Agent using PDF documents, spreadsheets, websites, and images.

<Info>
  **Prerequisites**

  * **Access:** Logged in to [AI Studio](https://app.bitbybit.studio/ai-studio).
  * **Supported Files:** PDF, CSV, XLSX, DOCX, TXT (Max 10MB).
  * **Supported Images:** JPG, PNG (Max 5MB).
</Info>

For a step-by-step guide, please visit our [**YouTube Channel**](https://www.youtube.com/@bitbybit.studio) or watch the video below:

<iframe width="560" height="315" src="https://www.youtube.com/embed/5ncDKiPHGMA" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; fullscreen" allowfullscreen />

## Why use diverse data sources?

Your AI Agent needs different types of "brain food" to be effective. While a **PDF** might contain your return policy, a **CSV** might hold your pricing list, and an **Image** acts as a visual menu to send to customers. Combining these sources creates a robust, expert agent.

## Data Source Types

Understand which source fits your content.

| Type        | Format                    | Best For                                                   |
| :---------- | :------------------------ | :--------------------------------------------------------- |
| **File**    | PDF, DOCX, TXT, CSV, XLSX | Internal SOPs, Product Manuals, Price Lists.               |
| **Text**    | Raw Text                  | Paste specific paragraphs or temporary rules directly.     |
| **Website** | URL                       | Dynamic content like blog posts or help centers.           |
| **Image**   | JPG, PNG                  | Visual assets the AI *sends* to users (e.g., Menus, Maps). |

## Configuration steps

<Steps>
  <Step title="Navigate to Knowledge Base">
    1. Open your AI Studio dashboard.
    2. Click on **Knowledge Base** in the side menu.
    3. Click the **Add Knowledge** button in the top right.

           <img src="https://mintcdn.com/asmaraku/WlT_6AuzhITpdgTf/images/Screenshot2025-12-23103440.png?fit=max&auto=format&n=WlT_6AuzhITpdgTf&q=85&s=41944065f33b6f294d3818ba7be2f189" alt="Screenshot2025 12 23103440" width="774" height="515" data-path="images/Screenshot2025-12-23103440.png" />
  </Step>

  <Step title="Select and Configure Source">
    Choose the tab that matches your data type.

    <AccordionGroup>
      <Accordion title="Option A: File (Docs & Data)">
        **Upload documents directly.**

        1. Select the **File** tab.
        2. Drag and drop your file (Max 10MB).
        3. Give it a clear Title (e.g., "Return Policy 2025").
        4. Click **Save knowledge**.
      </Accordion>

      <Accordion title="Option B: Website (Crawl)">
        **Connect your live web content.**

        1. Select the **Website** tab.
        2. Enter the **Website URL**.
        3. Choose your crawl method below.

        **Method 1: Single Page (Default)**

        * **What it does:** Indexes *only* the specific article or page you provide.
        * **When to use:** For a specific blog post, product page, or news article.
        * **Example:** `https://yourwebsite.com/blog/latest-feature` (Unchecked box).

        **Method 2: Crawl Entire Website**

        * **What it does:** Looks for the sitemap and systematically indexes *all* connected pages.
        * **When to use:** To teach the AI about your entire company, products, and services.
        * **Example:** `https://yourwebsite.com` (Checked box).

        <Tip>
          **⚠️ Deep Dive: Crawling Best Practices**

          For "Crawl entire website" to work optimally, you **must provide the main domain**.

          * **✅ Do this:** Enter a root domain like `https://bitbybit.studio`. This allows the crawler to easily find the sitemap.
          * **❌ Avoid this:** Do not enter a specific sub-page like `https://bitbybit.studio/bitchat` with the crawl box checked. The crawler may fail to locate the sitemap from a deep link.
        </Tip>

        **Quick Decision Guide:**

        | Goal                         | URL Type                  | Crawl Option   |
        | :--------------------------- | :------------------------ | :------------- |
        | Teach AI one specific post   | `https://myblog.com/post` | **Unchecked**  |
        | Teach AI entire company info | `https://mycompany.com`   | **Checked** 👍 |
      </Accordion>

      <Accordion title="Option C: Image (Visual Assets)">
        **Teach the AI to send images.**

        1. Select the **Image** tab.
        2. Upload your JPG/PNG.
        3. **Critical:** In the "When should the AI send this image?" field, type the trigger prompt (e.g., "Send this when user asks for the dinner menu").
      </Accordion>
    </AccordionGroup>
  </Step>
</Steps>

## Next steps

Once your data is uploaded, verify that the agent is reading it correctly.

<CardGroup cols={2}>
  <Card title="Test in Playground" icon="flask" href="#https://app.bitbybit.studio/ai-studio/ai-agent/">
    Ask questions related to your uploaded PDF or Website to test accuracy.
  </Card>

  <Card title="Configure Skillsets" icon="wrench" href="https://knowledge.bitbybit.studio/ai-studio/how-to-setup-skillset-in-ai-studio">
    Enable "Product Recommendations" if you uploaded a product catalog CSV.
  </Card>
</CardGroup>