The Staffbase MCP Server allows
MCP Server tools
MCP Server supports the following read-only tools:
| Tool | Purpose | Returns |
|---|---|---|
who_am_i |
Identify the authenticated user. | The profile information of the signed-in user. |
search |
Search Staffbase content. | Matching News, Pages, and User Profiles that the user can access. |
find_media |
Find media content. | Media items that the user can access. |
get_media_download_url |
Retrieve the download URL for a media item. | A download URL for media the user can access. |
inspect_media |
Retrieve metadata for a media item. | Metadata for a media item the user can access, such as format and size. |
get_page_content |
Retrieve the full content of a page by ID. | Full page content for a page the user can access. Returns plaintext by default and can return HTML when page structure matters. |
list_news_posts |
List published news posts, newest first. | A paginated list of summaries (IDs, basic metadata, localized titles and teasers) of news posts that the user can read, plus a cursor for the next page. Supports optional filtering by text query and content type (articles, updates, pictures). Does not include comment, like, or other engagement counts. |
get_news_post_content |
Retrieve the full content of a news post by ID. | The post ID, basic metadata, and content for each locale (title, teaser, HTML body, and media fields) of a news post that the user can read. |
search_emails |
Search Staffbase emails. | Drafts, scheduled, and sent emails matching by title, date, or author that the user can access. |
get_email_content |
Retrieve the full content of an email. | The full content of an email in all localized versions that the user can access. |
search returns results ranked by relevance and provides content snippets only. Use get_page_content to retrieve full page content from a specific page, or get_news_post_content to retrieve the full content of a news post.
Troubleshooting
Use the following table to diagnose common connection and authentication issues.
| Issue | Possible Cause | Resolution |
|---|---|---|
Authorization fails with 401 or 400. |
The managed OAuth client is missing, or is misconfigured. | Verify that your OAuth client matches the guide for your AI client. |
| Authorization flow does not start. | OAuth is not configured. | Ensure OAuth is activated for your platform and configure the OAuth client access. |
| Login succeeds, but the redirect fails. | The AI client’s callback URL is not registered. | Add the exact redirect URI to the OAuth client configuration. |
| The browser client reports a CORS error. | The client origin is not allowlisted. | Add the origin to Allowed CORS Origins in the OAuth client configuration. |
| Search returns few or no results. | Search results are limited to the authenticated user’s permissions and currently include only news, pages, and user profile information. | Sign in as a user with access to the content you expect. |