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

# Train on files and connected apps

> Upload documents directly or sync content from Notion, Google Drive, Dropbox, OneDrive, Box, and GitHub.

**Files & Data Sources** is where your chatbot learns from documents. You can
upload files by hand or connect an app and sync files from it.

## Upload files directly

1. In your chatbot's sidebar, open **Files & Data Sources**.
2. Select **Add Files** → **Local Files**.
3. Choose your files with **Browse File** and confirm with **Upload Files**.

Uploads accept common document formats (PDF, DOC/DOCX, TXT, CSV, and similar
text-based formats) up to **10 MB per file** by default. If you need larger
files, contact [support@sitegpt.ai](mailto:support@sitegpt.ai).

<Tip>
  The chatbot learns from the extractable text. Scanned PDFs that are pure
  images, and audio or video files, will not produce useful training content.
</Tip>

## Connect an app

Connected sources let you pick files from the tools where your content
already lives, and re-sync them when they change:

<CardGroup cols={3}>
  <Card title="Notion" icon="file-lines" href="/docs/data-sources/notion" />

  <Card title="Google Drive" icon="google-drive" href="/docs/data-sources/google-drive" />

  <Card title="Dropbox" icon="dropbox" href="/docs/data-sources/dropbox" />

  <Card title="OneDrive" icon="microsoft" href="/docs/data-sources/onedrive" />

  <Card title="Box" icon="box" href="/docs/data-sources/box" />

  <Card title="GitHub" icon="github" href="/docs/data-sources/github" />
</CardGroup>

The connectors follow the same flow (GitHub differs slightly: you point
it at a repository and file patterns instead of picking files):

<Steps>
  <Step title="Connect your account">
    Open **Files & Data Sources** → **Add Files**, pick the app, and select
    **Connect Account**. You will be sent to the provider to authorize
    read access.
  </Step>

  <Step title="Choose what to sync">
    Browse or search your files, select the ones the chatbot should learn
    from, and confirm with **Sync Selected Files**.
  </Step>

  <Step title="Watch training complete">
    Synced files appear in the list with live **Trained / Pending / Failed**
    status badges, exactly like uploaded files.
  </Step>
</Steps>

<Note>
  Disconnecting an app later does not delete already-trained content. Remove
  the documents from the list if you also want the chatbot to stop using them.
</Note>

## Editing a document's content

Open any document in the list to view its parsed text. For text-based
documents you can edit the parsed content directly and save, which retrains
the chatbot on your edited version. This is handy for fixing extraction
glitches without touching the original file.

## Working with the list

The Files & Data Sources list has the same tools as Website Links: live
status counts, search, bulk **Resync** and **Delete**, and CSV export. See
[managing your content](/docs/content/overview#managing-your-content).

<Info>
  Using the [CLI or API](/docs/developers/overview)? SharePoint and Confluence
  connectors are available there in addition to the sources above.
</Info>