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

# Adding knowledge

> Create pages and folders, upload documents, and import links into the Knowledge Base.

There are four ways to put content into Knowledge: write a page, create a folder to organize pages, upload a document, or import a link. Everything you add lands in the folder tree described in the [Knowledge Base overview](/build-your-agent/knowledge-base/overview).

## Add An Item

<Steps>
  <Step title="Open Knowledge">
    Go to **Agents > Knowledge**.
  </Step>

  <Step title="Choose what to add">
    Click the add button and choose **Page**, **Folder**, **Link**, or **Document**.
  </Step>

  <Step title="Choose a destination">
    Put the new item in the folder where the agent should expect to find that topic.
  </Step>

  <Step title="Review the result">
    Pages are available as soon as you save. Documents and links need to finish processing first.
  </Step>
</Steps>

## Pages And Folders

Write a page when your team owns the content and needs to edit it directly. Create a folder when you want to group pages and imported files by topic, region, listing, product, or team.

Keep each page focused on one topic. Broad pages that mix several policies retrieve less accurately than short, single-purpose pages. See [Knowledge writing principles](/build-your-agent/what-goes-where/knowledge) for how to structure the content itself.

## Documents

Document uploads support these file types:

* PDF
* DOCX
* CSV
* XLS
* XLSX
* JSON
* MD
* TXT

Each uploaded file can be up to 25MB. Markdown and text files can be imported as inline content. Other file types are uploaded, processed, and attached to source-backed Knowledge nodes.

<Tip>
  Structured tabular data such as a cleaning schedule or vendor list is a natural fit for a CSV, XLS, or XLSX upload. Import the file instead of retyping it as a page.
</Tip>

## Links

Links can be imported in two modes:

| Mode | Use it for |
| - | - |
| **Individual URL** | Import one page. |
| **Crawl** | Discover pages from a site, choose the pages to import, and place them in a Knowledge folder. |

A crawl can import up to 600 pages at a time. Failed link imports can be retried from the source list.

<Note>
  Documents and links become source-backed Knowledge, which means their processing state is tracked and they can be re-synced. See [Connected Sources](/build-your-agent/knowledge-base/connected-sources) for status meanings and recovery steps.
</Note>

## Choosing Between An Upload And A Connection

| Situation | Use |
| - | - |
| A one-off file that will not change often | Document upload |
| A single web page or help centre section | Link import |
| Content that lives in Google Drive or Notion and changes regularly | A [connected source](/build-your-agent/knowledge-base/connected-sources) |
| Listing, reservation, or property data from a PMS | The [PMS integration](/build-your-agent/knowledge-base/pms-knowledge) |

## Frequently Asked Questions

<AccordionGroup>
  <Accordion title="What file types and sizes can I upload?">
    PDF, DOCX, CSV, XLS, XLSX, JSON, MD, and TXT, up to 25MB per file. Markdown and text files can come in as inline content.
  </Accordion>

  <Accordion title="How many pages can a website crawl import?">
    Up to 600 at a time. You choose which discovered pages to import and which Knowledge folder they land in.
  </Accordion>

  <Accordion title="Should I upload a file or connect the source system?">
    Upload when the content rarely changes. Connect the source when it changes regularly and you want scheduled syncing to keep Knowledge current without a re-import.
  </Accordion>

  <Accordion title="Can I import a spreadsheet instead of retyping it as a page?">
    Yes, and you should. Schedules, assignment tables, and vendor lists import cleanly as CSV, XLS, or XLSX.
  </Accordion>
</AccordionGroup>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.