Just ask Payrun to
No dashboards to click through. Connect Claude — or any MCP client — and talk to your HR data directly. Timesheets, attendance, leave history, and your org directory — all one plain-language request away.
01 — CONNECT
One connection, every client.
Add the Payrun MCP endpoint to Claude, Cursor, or any MCP-compatible agent. No custom integration, no glue code — it speaks the protocol natively.
MCP-compatible agent
MCP Server
Payrun Core
team_timesheet
leave_records
attendance_overview
list_employees
flagged_hours_summary
report_by_employee
02 — TOOLS
A real toolbox, not a chatbot.
Dozens of scoped tools covering timesheets, attendance, leave, and your org directory — each one a precise read your agent can pull, not a guess.
03 — AUDIT
Every call, on the record.
Agent actions land in the exact same activity log as your team's. Know who — or what — pulled a timesheet at 9:14am.
09:14:02 agent:claude team_timesheet(dept=design) OK
09:14:11 agent:claude leave_records(emp=88) OK
09:15:40 human:jonas attendance_summary(range=week) OK
04 — CONTROL
Locked to what you allow.
Grant read-only access to timesheets, attendance, and leave data — you decide per role, per tool, per organization.
SECURITY MODEL
Scoped tokens. Real audit trail.
Every MCP session runs as an OAuth-scoped, read-only Payrun role — never a superuser. Agents only see the reporting tools you grant, and every call lands in the same audit log as a human click.
OAuth 2.1 scoped sessions
Per-tool permission control
Immutable audit log
Frequently Asked Questions
What is MCP?
The Model Context Protocol is an open standard that lets AI assistants like Claude connect securely to external tools and data sources. Payrun's MCP server exposes HR and payroll actions as tools an agent can call.
How do I connect Payrun to my AI assistant?
The Payrun MCP server is live. Add Payrun as a connector in your AI assistant, sign in with your Payrun account and approve access. Your assistant can then read the data your role allows.
Which AI clients does it work with?
Any MCP-compatible client, including Claude, ChatGPT and Grok, with no Payrun-specific plugin required.
Can I limit what an agent can do?
Yes. Access is scoped per role and per tool — you choose exactly which actions, like approving leave or running payroll, an agent is allowed to perform.
Is every action logged?
Every MCP tool call is recorded in the same audit trail as actions taken by your team in the Payrun app, so you always know who — or what — did what.