hireEZ MCP: Connect Your ATS to Claude and ChatGPT
—— Read time: 6 min
Your pipeline just joined the conversation.
Bring hireEZ into Claude or ChatGPT and work your reqs in plain language — search your Talent Pool, read pipeline health, move candidates, leave notes. No exports. No copies. No new password. The assistant acts as you, with exactly the access you already have.
What is hireEZ MCP?
hireEZ MCP is a secure connection that brings your hireEZ Talent Pool and recruiting pipeline into AI assistants like Claude and ChatGPT. Recruiters can search candidates, read pipeline health, move people between stages, and leave notes — in plain language, inside the assistant they already work in.
It closes the gap between where recruiters think and where the work lives without handing your candidate data to an AI tool wholesale. Each person signs in as themselves and inherits exactly their own hireEZ permissions; candidate contact details are never returned; and nothing can be deleted. It's available for Claude and ChatGPT, in controlled release, and built on the Model Context Protocol — the open standard for connecting AI assistants to business systems.
The gap it closes
Recruiters live in two places at once: the assistant where they think, and the ATS where the work is.
Every answer means leaving the conversation, finding the req, reading the stage, coming back.
hireEZ MCP closes that gap — without handing your candidate data to an AI tool wholesale.
See it in the chat
Say it in the conversation. It's done in hireEZ.
"Which candidates are stuck at onsite for the Platform Engineer req?" Four have been at Onsite more than 10 days. Priya Raman is the longest, at 18.
"Summarize Priya's background for the role." Eight years, distributed systems, two scale-up exits. Strongest on the infra requirement. No direct fintech experience.
"Move her to Offer and note that the panel was unanimous." Moved to Offer. Note added.
Illustrative example. Names and details are invented.
Three questions. No tabs. Nothing exported.
What your team can do
Find people you already have
Brief it like you'd brief a sourcer — against your own Talent Pool. That candidate you know is in there somewhere, from a req two quarters ago? Search by skills, titles, companies, locations, tags, and years of experience, and pull the full profile of anyone in it — without opening hireEZ to go looking.
Run the pipeline from the chat
See who's on a req, at what stage, and move them. The stage update you meant to do after the debrief and didn't. List the stages a project offers, see where every candidate sits, and move them forward — in the same conversation where you just decided to.
Read the room in one question
Pipeline health without building a view. "Which reqs are at risk?" used to be a dashboard you had to go find. Ask it instead — where candidates are piling up, where they're going cold, which stage is quietly eating the week.
Keep it clean as you go
Notes and tags at the moment you think of them, not at the end of the day. The context that normally dies in your head between the call and the CRM. Add a note, add or remove tags, find a project by name — the admin tax, paid in a sentence.
Admins can also manage workspace members and user groups, and invite new teammates.
Why it's safe to turn on
This is the section your security team will actually read. All of it is true in the product today. For hireEZ's certifications and data practices, see our Information Security and Compliance page.
It acts as the person asking. Never more.
Each user signs in with their own hireEZ account through your normal SSO. The assistant inherits exactly their permissions, in exactly their workspaces — checked live, on every single request. No service accounts. No shared keys. No team-wide token.
Candidate contact details are never returned. To anyone.
Email addresses and phone numbers are not available through hireEZ MCP under any configuration, for any user, on any plan. The assistant can tell you a record has contact details, so you know to open hireEZ. That's the whole of it.
Nothing can be deleted.
There is no delete capability of any kind — not candidates, not projects, not notes, not users. No bulk export either. The strongest thing an assistant can do is move a candidate to another stage.
You choose what's live.
Turn individual capabilities on or off for your organization, or for a single workspace. Most teams start read-only while they evaluate and switch on writes once they trust it. Changes land within about 15 minutes — nobody reconnects, nobody re-authorizes.
Access ends when employment does.
Remove someone from hireEZ and their assistant loses access automatically. Access tokens last 15 minutes, and every renewal re-checks that the person, their workspaces, and your organization's access are all still valid.
Only approved AI tools can connect.
Most MCP servers let any application register itself and start asking for access. hireEZ doesn't. Every AI client is verified against an approved list before it can even show a sign-in screen to your people.
What is the Model Context Protocol?
MCP — the Model Context Protocol — is the open standard for connecting AI assistants to business systems.
Building on it means hireEZ isn't a one-off integration with a single vendor. As your company adopts new AI tools, the same governed connection works with them, under the same controls. Support for additional assistants is planned.
Underneath: OAuth 2.1 with PKCE over encrypted transport — the same authorization model as your other enterprise SaaS integrations.
How to set it up
Live in about two minutes, once access is enabled.
Ask us to switch it on. hireEZ MCP is in controlled release. Your Account Manager enables it for your organization.
Add the connector. One URL in your AI tool's connector settings.
Everyone signs in as themselves. Each person connects with their own hireEZ account and gets their own permissions.
Frequently asked questions
Which AI assistants are supported?
Claude (claude.ai, Claude Desktop and Claude Code) and ChatGPT, in controlled release. Any assistant that implements MCP with per-user OAuth is technically compatible, and hireEZ approves new clients on request. Ask your Account Manager for the current list.
Does hireEZ see our AI conversations?
No. hireEZ receives structured requests — which capability, with which parameters. Never the prompts your team types, never the assistant's replies.
Is our candidate data used to train an AI model?
Not by hireEZ. Data returned to your AI assistant is governed by your agreement with that vendor, which you hold directly.
Can the AI act on data a user can't see?
No. Every request is authorized against that person's live hireEZ permissions, in the workspace it targets.
Can we start read-only?
Yes, and most customers do. Search and reporting first. Writes when you're ready.
Can candidate contact details ever be exposed through the assistant?
No. Email addresses and phone numbers are never returned through hireEZ MCP, under any configuration, for any user, on any plan. The assistant can only tell you that a record has contact details, so you know to open hireEZ.
Can the assistant delete anything?
No. There is no delete capability of any kind — not candidates, projects, notes, or users — and no bulk export. The most an assistant can do is move a candidate to another stage.
How is access revoked?
Remove the person from hireEZ and their assistant loses access automatically. Access tokens last 15 minutes, and every renewal re-checks that the user, their workspaces, and your organization's access are still valid.
What does it cost, and is it available on our plan?
hireEZ pricing is explained on our pricing page. hireEZ MCP is in controlled release, so your Account Manager can confirm whether it's available on your plan.
Get a walkthrough
Ask. Approve. It's in hireEZ.
Fifteen minutes with our team, on your reqs, with your data. Not a canned demo.
