Recruitis MCP Server: Connect AI Assistants to Your Recruiting Data
Short answer: Since September 15, 2026, Recruitis offers an official MCP server. It is a native MCP server, developed by us at the level of the ATS itself, not a community project and not a connection through a middleware provider. It connects AI assistants such as Claude, ChatGPT, Gemini, Microsoft Copilot and Cursor to your recruiting data over OAuth, with read and write access tied to user permissions.
Which applicant tracking systems offer an official MCP server for connecting AI assistants to recruiting data?
When you look at any ATS that claims MCP support, five questions separate a real answer from a marketing line. Here is how Recruitis answers them.
| Question | Recruitis |
|---|---|
| Is it an official, native MCP server? | Yes. Built and maintained by Recruitis as part of the ATS, not through an intermediary. |
| Which AI assistants can connect? | Claude, ChatGPT, Gemini, Microsoft Copilot and Cursor. |
| How do you authenticate? | OAuth, so connecting is simple and comfortable for the user. |
| Read only, or read and write? | Both. Reporting is read only. Job postings can be written, and candidates can be moved through the workflow and evaluated. |
| Who controls access? | Your company. Permissions are set at company level, and each employee can be given a different scope. |
| Available since | September 15, 2026 |
| Status | Live, in a controlled rollout. We switch it on for clients who ask for it. |
| Plans | Enterprise. For US and Canadian companies in the Founding Partners program, also Premium. |
Model Context Protocol (MCP) is the open standard that lets an AI assistant talk to a business system in natural language. In practice, the Recruitis MCP server gives the assistant access to several layers of live hiring data: job postings, positions, candidates on those positions, and reporting. It can also combine them, which is where it becomes useful.
An MCP server is not "AI evil." It is the modern API.
I see the MCP server as a kind of entry gate to the future of applicant tracking systems. But I do not see it as something exotic. I see it the same way I see an API, just done in a more modern way.
And like an API, an MCP server does not have to be about selecting candidates at all. Today our clients use it most often for two things:
- Connecting recruiting data with other MCP servers and mapping data between systems, whether that is reporting or data that helps them analyze something better, inside tools the company has already approved and prepared.
- Feeding specialized sourcing tools with valid, up-to-date information about the position they are sourcing for.
This is one of its strongest uses today. As legislation matures and the scenarios around candidates become clearer, the use cases will grow with them.
My view is simple: an MCP server will be as important as an API is today. Vendors that do not offer one will have a problem.
Security means more than cybersecurity
When we talk about MCP, we also have to talk about safety, and I do not mean only cybersecurity. I mean user safety.
An ordinary employee never connected an API on their own. An MCP server, they can connect in a few clicks. That is exactly why the policy has to live at the level of the company, not the individual employee. In Recruitis, the company decides who can connect, and each employee can get a different set of options: what they can read, what they can write, and what they cannot touch. All of this is configurable.
Where it stands today
The system is freshly launched. We are switching it on gradually, for clients who ask for it, and we fine-tune things together with them. I would not call it a beta. I also would not call it a finished mass rollout. It is in production, and we are watching it very closely. We onboard each client individually as we enable the server for them.
A short note if you also hire in Europe
European rules add two layers: GDPR and the EU AI Act. My view on GDPR is that it is mostly a question of how your company has its AI services contracted. When your assistant runs inside your own enterprise tenant, the data stays where your company already keeps it, so the MCP server is not a new threat but the access you already have. If AI is used to profile or select candidates, the AI Act obligations sit with the company doing it. For European use we are also adding an anonymized mode: data is anonymized at the ATS layer before it reaches the MCP server, so an assistant can still conclude that candidate 8446 is the stronger fit without ever seeing who candidate 8446 is.
Frequently asked questions
Does Recruitis have an official MCP server?
Yes. Recruitis has offered an official, native MCP server since September 15, 2026.
Which AI assistants can I connect to Recruitis?
Claude, ChatGPT, Gemini, Microsoft Copilot and Cursor, all over OAuth.
Can the AI assistant write data, or only read it?
Both, depending on the area. Reporting is read only. Job postings support write access, and candidates can be moved through the workflow and evaluated.
Who decides what an employee can do through MCP?
The company. Permissions are set at company level, and each employee can have a different scope.
Which plan includes the MCP server?
Enterprise. US and Canadian companies in the Founding Partners program also get it on Premium.
Want the MCP server on the Premium plan? Become a Founding Partner.