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

# Files

> Upload PDFs, Word documents, text and markdown

## Supported files

Upload **PDF**, **Word (.docx)**, **text**, **markdown**, **CSV**, **HTML** or **JSON**, up to **25 MB** each. You can select several at once or drag them onto the upload box.

Clarky converts each file to plain text and creates one document per file. The text is fully editable afterwards, so you can fix anything the conversion got wrong or trim sections you do not want the agent to use.

<Note>
  Scanned PDFs (images of pages with no text layer) cannot be read. Run them through OCR first, or paste the text into a document instead.
</Note>

## Uploading

<Steps>
  <Step title="Add > Upload files">
    From **Context > Documents**.
  </Step>

  <Step title="Drop your files">
    Or click to browse. Remove any file from the list before uploading with the ×.
  </Step>

  <Step title="Choose channels">
    Applied to every file in this batch. You can change channels per document later.
  </Step>

  <Step title="Upload">
    Each file is converted and indexed. Files that could not be read stay in the list with the reason so you can fix and retry.
  </Step>
</Steps>

## Tips

* Long files are usually best **retrieved** (Always include off). Pinning a 40 page manual sends the whole thing with every message.
* Files checked for **Voice** count against the voice budget in full. Leave long files off Voice.
* Uploading a new version of a file creates a new document; delete the old one when you are done.