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

# Train Your Agent on Files, Websites, and Text Content

> Add PDF, DOCX, TXT, XLSX, and CSV files, website URLs, or plain text as training sources so your BotSmith agent answers questions accurately.

BotSmith agents use **Retrieval-Augmented Generation (RAG)** to answer questions. Instead of relying on the model's general knowledge, your agent retrieves the most relevant passages from your own content and uses them to construct each response. This keeps answers accurate, grounded, and specific to your business.

You add content to an agent as **training sources** — files, websites, or raw text — and BotSmith indexes them automatically in the background. The more relevant content you add, the better your agent answers.

## Supported Source Types

<CardGroup cols={3}>
  <Card title="File Upload" icon="file-arrow-up">
    Upload documents directly from your computer. Supported formats: **PDF, DOCX, TXT, XLSX, CSV**. Maximum file size: **100 MB**.
  </Card>

  <Card title="Website Scraping" icon="globe">
    Provide a URL and BotSmith fetches and indexes the visible text from that page. Ideal for product pages, help centers, and blog posts.
  </Card>

  <Card title="Plain Text" icon="align-left">
    Paste any text content directly — FAQs, policy excerpts, internal notes — and give it a name so it's easy to identify later.
  </Card>
</CardGroup>

## Uploading a File

<Steps>
  <Step title="Open your agent">
    From your dashboard, click **Manage** on the agent you want to train.
  </Step>

  <Step title="Go to the Sources tab">
    Click the **Sources** tab inside the agent builder.
  </Step>

  <Step title="Open the Add Source modal">
    Click **Add Source**.
  </Step>

  <Step title="Select the File tab">
    Choose **File Upload** from the modal tabs.
  </Step>

  <Step title="Choose your file">
    Click the upload area and select a file, or drag and drop it onto the area.
  </Step>

  <Step title="Upload">
    Click **Upload File**.
  </Step>
</Steps>

BotSmith reads your file, chunks it into searchable passages, generates embeddings, and stores them in your agent's vector index. The source card in the Sources tab shows a **Processing** badge while this happens and a **Completed** badge when it's ready.

**Supported formats:** PDF, DOCX, TXT, XLSX, CSV\
**Maximum file size:** 100 MB per file

## Adding a Website

<Steps>
  <Step title="Open your agent">
    From your dashboard, click **Manage** on the agent you want to train.
  </Step>

  <Step title="Go to the Sources tab">
    Click the **Sources** tab inside the agent builder.
  </Step>

  <Step title="Open the Add Source modal">
    Click **Add Source**.
  </Step>

  <Step title="Select the Website URL tab">
    Choose **Website URL** from the modal tabs.
  </Step>

  <Step title="Enter the URL">
    Paste the full URL (e.g., `https://example.com/help`) into the field.
  </Step>

  <Step title="Add">
    Click **Add Website**.
  </Step>
</Steps>

BotSmith fetches the page, strips navigation and other non-content elements, and indexes the visible text. Processing runs in the background — the source card shows a **Processing** badge until it completes.

For best results, add individual pages rather than entire domains. If your help center has a dedicated URL per article, add each article as its own source.

## Adding Plain Text

<Steps>
  <Step title="Open your agent">
    From your dashboard, click **Manage** on the agent you want to train.
  </Step>

  <Step title="Go to the Sources tab">
    Click the **Sources** tab inside the agent builder.
  </Step>

  <Step title="Open the Add Source modal">
    Click **Add Source**.
  </Step>

  <Step title="Select the Text Content tab">
    Choose **Text Content** from the modal tabs.
  </Step>

  <Step title="Enter a name and paste content">
    Enter a descriptive **Name** (e.g., "Refund Policy Q3 2025") and paste your content into the text area.
  </Step>

  <Step title="Save">
    Click **Add Text Content**.
  </Step>
</Steps>

Plain text sources are processed immediately — the source card shows **Completed** straight away.

## Managing Sources

All sources for an agent appear in the **Sources** tab. Each card shows:

* **Source name** and type icon
* **File size** (for uploaded files)
* **Status**: Processing, Completed, or Failed
* **Date added**

To **delete a source**, click the trash icon on its card and confirm the deletion in the modal.

<Warning>
  Deleting a source removes it from the agent's knowledge base immediately. The agent will no longer use that content to answer questions. This action cannot be undone.
</Warning>

After you add or delete a source, re-indexing happens automatically — you don't need to manually retrain or redeploy the agent.

## Source Limits

Your plan determines how many file uploads, website sources, and text sources you can add across all agents. You can monitor your usage on the dashboard's **Resource Usage** panel.

<Note>
  Plan limits apply separately to file uploads, website sources, and text sources. If you reach a limit, you'll need to delete an existing source or upgrade your plan. View and manage your plan at [Account & Plans](/account/plans).
</Note>