> ## Documentation Index
> Fetch the complete documentation index at: https://support.i.moneyforward.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Supported knowledge integrations

> Feature support matrix for the knowledge sources that work with AI Helpdesk (Confluence, SharePoint, Google Drive, Notion, Box, Admina MCP).

AI Helpdesk connects to your company's document services as **knowledge sources**. Once connected, the AI agent can search and reference documents in these services when answering inquiries.

Documents imported from integrations are **read-only** in Admina. Edit them in the source service.

## Feature support

| Integration | Search and reference documents | Reading image content | Attaching images to answers | Connection method |
| - | :-: | :-: | :-: | - |
| [Manually added knowledge](/en/ai-helpdesk/knowledge/adding) | Yes | Partial \*3 | Yes | Upload directly in Admina |
| Confluence | Yes | Yes \*1 | No | OAuth or API token |
| SharePoint (including OneDrive) | Yes | Yes \*1 | No | Microsoft admin consent |
| Google Drive | Yes | Yes \*1 | No | Service account (domain-wide delegation) |
| Notion | Yes | Partial \*3 | Yes | OAuth |
| Box | Yes | Yes \*1 | No | OAuth |
| Admina MCP server | — \*2 | No | No | MCP (tool integration) |

\*1 Images inside documents are imported as descriptive caption text. The AI references the converted text, not the images themselves.

\*2 The Admina MCP server is not a document search integration. It gives the AI agent tools to view and manage IT asset information.

\*3 Images are imported as alt text, which the AI can reference. The image content itself (figures or text in the image) is not read.

<Note>
  * Reading image content — whether the AI can use the content of images inside documents as a source for its answers. On Confluence, SharePoint, Google Drive, and Box, images are converted to descriptive text at import, so their content can be referenced (the images themselves are not shown). For manually added knowledge, only the alt text is referenced.
  * Attaching images to answers — whether the Bot can attach images contained in documents to its answers. Among the external knowledge sources, only **Notion** is supported.
</Note>

For details such as supported file formats and sync scope, see each integration's page.

## Setup and specifications per integration

* [Manually added knowledge](/en/ai-helpdesk/knowledge/adding)
* [Confluence](/en/ai-helpdesk/integrations/confluence)
* [SharePoint](/en/ai-helpdesk/integrations/sharepoint)
* [Google Drive](/en/ai-helpdesk/integrations/google-drive)
* [Notion](/en/ai-helpdesk/integrations/notion)
* [Box](/en/ai-helpdesk/integrations/box)
* [Admina MCP server](/en/ai-helpdesk/integrations/admina)

## Related pages

* [AI Helpdesk integrations overview](/en/ai-helpdesk/integrations/index)
* [Add knowledge](/en/ai-helpdesk/knowledge/adding)


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