Warsaw, May 28, 2026 - Calamari is shipping an MCP server (Model Context Protocol) that lets you connect your AI assistant directly to your HR system. The feature is rolling out to all customers now.
Why this matters
ChatGPT, Claude, Copilot, Gemini - more and more people spend their workday inside these tools. Writing, analyzing data, planning. But to check who's out, request time off, or clock in, you still had to open Calamari in a separate tab.
The MCP server fixes that. Instead of switching between apps, you tell your AI assistant what you need. It connects to Calamari and does it for you.
What you can do today
Once Calamari is connected to your AI assistant, you can:
- Clock in and clock out - without opening the app
- Submit a time off request - with a substitute, reason, and comment
- Check upcoming absences on your team
- Search the employee directory
Type "Clock me in," "Request time off for May 19," or "Who's out today?" Your assistant connects to Calamari and handles it.
How it works
First, your administrator enables the AI integration in company settings (Configuration > Integrations > AI Integration). Once it's on, every employee can connect their own AI assistant. In your Calamari profile (AI Integration tab), you'll find step-by-step instructions for your specific tool. Step-by-step setup instructions are available for ChatGPT, Claude, Claude Code, Cursor, VS Code, Codex, Windsurf, Copilot Studio, and Gemini CLI, but any MCP-compatible client will work.
Depending on the tool, you copy the server URL, paste a config snippet, or run a single command. Sign in with your Calamari account through OAuth and you're set - you see exactly what your permissions in the system allow.
"Our customers have been using the Calamari API for years. MCP is the next step - the same access to data, but you don't need a developer. Just an AI assistant and a single question," says Kamil Wojewoda, Cofounder of Calamari.
Security and control
Every MCP request follows the same permissions as Calamari itself. You only see and modify data you have access to. Administrators decide who can use the integration and can turn it off at any time.
What's next
Calamari is actively building out the MCP server. Coming updates will add full employee data, timesheets, and leave entitlements. This will let managers generate custom reports and combine HR data with information from other apps, all through their AI assistant.
FAQ: Calamari launches MCP server - your AI assistant now talks to your HR system
What is MCP?
Model Context Protocol is an open standard that lets AI tools (like ChatGPT or Claude) connect to external applications. It allows your AI assistant to perform actions in Calamari without opening a browser.
What is an MCP server?
An MCP server is a program that exposes a specific application's data and actions in a format AI assistants understand. It works as a bridge: it receives a request from your AI assistant, performs the operation in the app, and returns the result. Without an MCP server, an AI assistant can only work with information you paste into it, it cannot interact with external systems on its own. The Calamari MCP server exposes specific HR actions: clocking in and out, submitting time off requests, checking who's out today. Your assistant connects to the server, and the server performs the action on behalf of the logged-in user, with full permission enforcement.
Does this cost extra?
No. The MCP server is available to all Calamari customers as part of their current subscription.
Can AI see all employee data?
No. Once connected, you see exactly what you see in Calamari. Your permissions determine what data you can access. Administrators have broader access, while regular employees only see their own data and their team's data.
Is it secure?
As secure as using Calamari through your browser. You sign in with your own account via OAuth, and every request follows the same access rules. Administrators can disable the integration at any time.
Do I need IT to set it up?
No. Your administrator enables the integration in company settings (Configuration > Integrations), and then each employee connects their AI assistant on their own. You'll find instructions in your Calamari profile (AI Integration tab). Setup takes a few minutes.
Which AI tools are supported?
ChatGPT, Claude, Claude Code, Cursor, VS Code, Codex, Windsurf, Copilot Studio, Gemini CLI, and any other MCP-compatible client.
