Papershift

Getting a token

1. The Papershift API is a paid add-on: ask Papershift sales or your customer success manager to activate it for the account.
2. Sign in, click the gear icon (bottom left) → Account Settings → API tab → Create New API Key.
3. A new key has read and write permissions by default; restrict it to read if you only want reporting, then copy it into PAPERSHIFT_API_TOKEN. Anyone holding it can read the whole roster, so treat it as a password.

The token travels as a query parameter, not a header: Papershift's public API expects ?api_token=… on every request. AnythingMCP appends it automatically (QUERY_AUTH), so it never appears in a tool's parameters and never has to be pasted into a prompt.

Shift plan vs worked time. Papershift keeps the two apart, and mixing them up is the usual source of wrong answers:

• papershift_list_shifts is the plan — which shifts are planned, where and when.
• papershift_list_working_sessions is the record — who actually clocked in and out. A question about payroll wants working sessions; a question about next week wants shifts.

Date ranges are range_start and range_end (YYYY-MM-DD). Shifts and absences need both; working sessions need range_start. Shifts also need a location_id: call papershift_list_locations first.

Paging is page (1-based) with a fixed 25 rows per page; there is no page-size parameter. Locations and working areas come back in full, unpaged. Responses carry the rows under the resource name, e.g. {"shifts": [...]}, plus next_page when there is more.

Language: Papershift requires an interface_language on every request; the connector sends de automatically.

Cloud reachability: app.papershift.com is public with a valid certificate.