Troubleshooting
This page covers the most common issues when connecting an AI client to Klarify through the Model Context Protocol.
The AI client cannot find any Klarify tools
If your AI client reports that no Klarify tools are available, the connector is not active.
- Confirm the endpoint matches the MCP server URL shown under Settings → AI Assistant → MCP connections. A typo or missing path segment causes the connection to fail silently.
- Confirm you completed the browser sign-in step. Until you sign in, no tools are exposed.
- Restart the AI client and re-open the conversation.
I see fewer tools than I expect
Klarify advertises 4 to 11 workflow tools depending on your role and content access. If you are signed in as a full member without administrator access, you will not see the management tools manage_person, manage_org_structure, or bulk_import.
To use administration actions, sign in as a user with an admin role (Org Admin, Super Admin, Account Manager, or Account Owner). The server still checks every action inside a management tool because those roles do not all have the same permissions. See how Klarify AI applies your permissions.
A tool call returns “permission denied” or similar
Tool registration is one layer of access control; content-level permissions are another. Even if a tool is registered for your session, the underlying record may require additional access.
For example, an Org Admin who has not been granted access to a specific document cannot read that document, even through the AI client. Folders themselves are visible to all active members; access is enforced per document. Grant the required document access in the Klarify app and try again.
My session expired or I need to sign in as a different user
Each person connects with their own Klarify account. Starting another connection does not sign out people who are already connected.
To switch users:
- Remove the Klarify connector from your AI client.
- Add the connector again with the MCP server URL from Klarify.
- Sign in as the new user when the browser opens.
A bulk CSV import partially failed
The bulk_import tool returns created, already-existing, and failed results per row. Common causes of row-level failures:
- A required column is missing or misspelled in the CSV header
- A referenced name (for example, a region, department, position, or team) cannot be resolved
- A value is invalid, such as a display name instead of an email in a position’s
assigned_employeecolumn
Fix the offending rows and re-run the import. Successfully imported rows are not duplicated, and rows that already exist are safely ignored.
See CSV bulk import for the required columns of each record type, or Import people from CSV for the in-app workflow.
Still stuck?
Contact Klarify support at support@klarify.biz with:
- The AI client you are using (Claude Desktop, Microsoft Copilot Studio, or other)
- The exact prompt or tool name that failed
- Your Klarify organization and role