HR MCP server
Your HR data in Claude and ChatGPT
Connect Claude, ChatGPT, Cursor or VS Code to Skailer. Each person signs in as themselves, and the client gets their permissions and no more.
Safe by design: what the MCP connector can do
HR data shouldn't reach AI as a pasted export
Someone exports a table and pastes it into the chat. By tomorrow it's stale, and it shows the whole table. An admin API key gives the AI client everyone's data.
Connect Claude, ChatGPT, Cursor or VS Code in minutes
Add the Skailer server address in your client, sign in to Skailer in the browser and choose what the client may do. Claude Team and Enterprise owners add the connector once; each member connects their own account.
Ask about people, leave and goals without leaving chat
Once connected, the client reads the Skailer data you may see and answers with it: who is off in a release week, which goals are at risk, how a review cycle is going. Nobody exports a file first.
Request leave and approve requests from the chat
If you allowed changes when connecting, the client can make the same short list of changes SK.AI makes. Time off is checked first: balance, what's left and who approves. A request that breaks policy isn't created.
Give each connection one person's access
The connector has no company-wide key. Each person signs in from their client and gets exactly what their role opens. Reading is always on; changing data is a separate permission. Removing the connector ends access.
- Sign-in per personeach person signs in through the browser, with no shared token
- Two permissionsviewing your data is always on, and changing it is optional
- Role-based scopethe same field and data permissions as in Skailer itself
- Disconnect any timeremove the connector, and the client stops working with Skailer
See every tool and whether it reads or writes
Reading works with the default permission; changes need the permission to create and update data. Recruiting, compensation, documents and settings aren't included.
| Area | Read | Write (needs permission) |
|---|---|---|
| People | Search, profile, find by email, get me | — |
| Things to do | Get my inbox | — |
| Goals | Search goals, get a goal | Goals, key results, check-ins |
| Time off | Balances, categories, calendars, request check | Create, cancel, approve, reject |
| Surveys | Search, results, answers, participants | — |
| Performance reviews | Search, results, stats, participant report | — |
| Recognition | Search recognitions | Give a recognition |
Questions you can ask from Claude or ChatGPT
SK.AI, MCP or API: pick what fits the job
Inside Skailer, SK.AI answers from the side panel. In Claude, ChatGPT, Cursor or VS Code, MCP does the same under the same permissions. Systems use the API and webhooks.
What is the Skailer HR MCP server?
The Skailer HR MCP server, also called the MCP connector, links AI clients such as Claude, ChatGPT, Cursor and VS Code to Skailer. Each person signs in through the browser, and the client gets their permissions. It reads employees, goals, time off, surveys, reviews, Things to do and recognitions. If allowed, it updates goals, requests, cancels, approves or declines time off, and gives recognitions. It has no delete actions and no recruiting data.
SkailerKnow what the client can see and change
The client works only with the data the signed-in person may see in Skailer, and changes data only if they allowed it. Choose the client and plan your company already approves for work data.
Three ways HR data gets into an AI chat
Pasted exports need no setup, but they show the whole file and go stale. Connectors on one integration key are live, and the key often sees more than any one user should.
| Capability | Exports pasted into chat | Integration-key connector | Skailer MCP |
|---|---|---|---|
| Data freshness | As of the export | Live | Live |
| Who the client acts as | Whoever has the file | The key's account, often admin | The person who connected |
| Permissions | None after export | The key's scope | The person's role in Skailer |
| Changing data | Not possible | Whatever the key allows | Goals, own leave, approvals, if allowed |
| Read-only option | Not applicable | Depends on the key's scope | A choice on the access screen |
| Ending access | Delete every copy | Revoke the key for everyone | Remove the connector in the client |
How to connect Skailer to your AI client
See how teams use Skailer
Skailer MCP: common questions
MCP (Model Context Protocol) is an open standard that lets AI clients such as Claude and ChatGPT connect to other systems. An HR MCP server is the HR system's side of that connection: Skailer's lets your client read your HR data and, if you allow it, change it.
Yes. Add Skailer as a custom connector on Claude Free, Pro and Max; on Team and Enterprise the owner adds it once for everyone. Claude Code connects too. Each person signs in with their own Skailer account and gets their own permissions.
Yes, through ChatGPT developer mode. On Business, Enterprise and Edu the client can read and change data; on Pro it only reads. Cursor and VS Code with GitHub Copilot connect the same way, by the server address.
Only if you allow changes when you connect. Then it can update goals and key results, request or cancel your own time off, approve or decline time off waiting for you, and give recognitions. It can't delete anything.
Yes. Leave changes switched off on the access screen when you connect, and the client only reads. On ChatGPT Pro the connection is read-only anyway. You can reconnect later to allow changes.
SK.AI is the assistant inside Skailer. MCP gives the same data and actions to the AI client you already use. Both follow your permissions. Recruiting, compensation and documents aren't available through MCP today.