New AI Agent can now build your knowledge base, connect channels and invite agents — all by chat Try it now
Help homeProductPricingBlog
Start freeLog in

Upload knowledge documents

Add reliable customer-service material to the knowledge base.

Updated 2026-10-07

Walk through the demoDemo data
Knowledge baseDocuments · 3Processing 1New folderNew documentDemo data
DocumentsFAQ pairs
All contentSearch by title or tag…Test knowledge
DocumentSizeTagsUpdatedStatusActions
Brazil shipping timelines—shippingbrazilUpdated Jun 3, 2026, 4:00 PMSearchable⋯
EU customs & duties—shippingeuUpdated Jun 4, 2026, 2:30 PMIndexing⋯
30-day refund policy—returnsrefundUpdated Jun 2, 2026, 8:00 PMSearchable⋯
Page 1 · 3 documents
Step 1 / 16
Upload files

Click New document to open the dialog: drag in files in formats the product currently supports, with optional tags.

After product guides, policies, and procedures are uploaded, YundaDesk processes them so managed AI can use the content in answers.

Prepare the files#

Remove outdated, duplicate, or conflicting versions. Do not include secrets or unnecessary customer data. Use clear headings, and keep each document focused on a stable topic.

Supported formats include DOC/DOCX/DOCM, PPT/PPS/POT/PPTX/PPTM/PPSX/PPSM, XLS/XLSX/XLSM/XLSB, ODT/ODS/ODP, RTF, EPUB, CSV, PDF, TXT, Markdown, JSON, YAML, and EML.

Upload steps#

  1. Open the knowledge document page.
  2. Select the file upload action.
  3. Choose a format currently supported by the product.
  4. Wait for processing to finish.
  5. Select Test knowledge, ask a real question from the document, and inspect its source.

Select material from GitHub#

If GitHub appears in New document, you can sync text material from a public repository without uploading the entire repository.

  1. Open New document → GitHub and enter the public repository URL. Leave the branch empty to read the default branch.
  2. Select Next to read the repository and open Select and filter. Expand folders and select the folders or files you need. Clicking a name also selects it. To include everything, select Entire repository.
  3. Optionally expand Path filters (optional) and enter include and exclude paths, separated by commas. Both * and ** are supported. For example, include help/** and exclude help/examples/** to keep demonstration material out of the knowledge base.
  4. Select at least one file type: .md, .markdown, .mdx, .txt, or .rst.
  5. Select Next to open Preview and import. Check the matched file count and list, then select Import and sync. Select Back to make changes; your selections and filters are retained. Open the preview again after editing.

A folder selection includes matching files below it, including new files in later syncs. An individual file selection syncs only that file. Path and file-type filters apply together, and exclusions take priority. Import is unavailable without a selection or a matching file.

The import button appears only when the preview has matching files. You can also select a completed step at the top to go back and edit. You cannot skip to import before completing the preview. If no files match, go back and adjust your selections or filters. Longer file lists scroll within the list while the bottom actions remain visible.

After syncing, material appears under its repository source in the document list. Find the source row in All content, inspect its sync status, and open the action menu on the right:

  • Sync now reads your selected material again. Wait for an active sync to finish.
  • Auto-sync enables or pauses future automatic syncs. The setting is retained after a page refresh.
  • Disconnect and keep documents stops syncing after confirmation. Imported documents can still be used in AI replies.
  • Delete connection and documents stops syncing after confirmation and moves this source's documents to Trash, excluding them from answers. The remote repository and other sources are not changed. You can restore documents from Trash if needed.

Older disconnected sources also offer deletion of their retained documents from the list. Deletion and disconnection are unavailable while syncing. If a rate-limit error appears, try again later instead of repeatedly clicking.

After upload#

A listed file is not necessarily searchable. AI can rely on it only after successful processing into the current knowledge version. If processing fails, inspect the reason before uploading the same file again.

Successfully processed documents show a Use in AI replies status tag. To keep a document without using it in answers, turn off Use in AI replies from the document menu or detail page. The original content and automatically generated FAQs stop participating in answers, but the document remains in the knowledge base and can be enabled again without another upload. FAQs disabled individually are not restored automatically.

Didn't find your answer?

Start free and try the real product, or reach out — humans are as online as the AI.