Quick answer: If you already run your day in Claude or ChatGPT, you can connect KineticRecruiter to it with one link and then ask your candidate database questions in plain language, without copying anything in by hand. The connection uses MCP (Model Context Protocol), an open standard that lets an AI assistant talk to an outside tool. Once it is connected, the assistant can search and read your candidates, jobs and pipelines, and, when you approve it, create and update records. It only ever sees what your Kinetic login is allowed to see, and it asks before it changes anything. The connector is included on every KineticRecruiter plan.
The problem: your assistant is flying blind
You have probably already tried to use Claude or ChatGPT for recruitment. You paste in a CV, ask for a summary, maybe get it to redraft a job ad. It is genuinely useful, right up until the moment the work touches your actual database.
Then the friction starts. The assistant cannot see your candidates, so you copy details into the chat window by hand. It cannot see the live role, so you retype the brief. It gives you a shortlist built from what you pasted, not from the thousands of people already sitting in your system. Every useful answer costs you five minutes of copy and paste first, and the moment you close the tab, none of it is in your ATS.
That is the gap the connector closes. Instead of feeding the assistant scraps, you give it a safe, read-and-write line into your real database.
What MCP is, in one plain sentence
MCP, the Model Context Protocol, is an open standard that lets an AI assistant like Claude or ChatGPT connect to an outside tool and use it directly (Anthropic's Model Context Protocol).
That is the whole idea. Before MCP, an assistant only knew what you typed into it. With MCP, it can reach into a connected system, your ATS, your email, your calendar, and act on live data. KineticRecruiter runs an MCP server, so any assistant that speaks the protocol can connect to your account. It was the first agency ATS you could do this with, but that is not really the point. The point is what it does once it is on.
What you can actually do once it is connected
This is where it stops being a novelty. The value is not "a chatbot that knows your database". It is that your ATS becomes one tool among the several your assistant already holds, so whole jobs happen in one conversation. Here are four patterns recruiters actually use.
1. Search and shortlist without leaving the chat
You type:
Find me five candidates for the Sydney fintech BA role and tell me why each one fits.
The assistant runs the match against that role in your database and comes back with real people from your own records, each with a score and a written reason: strong on payments domain, four years as a BA, based in Sydney, open to contract. Not a generic AI answer, your candidates, ranked, with the reasoning shown. If you want to go wider, "show me anyone with insurance BA experience we haven't contacted in six months" works the same way.
2. Hand it a job brief and let it do the matching
A client emails you a two-page brief. Instead of reading it, keying a new role into your ATS, and then running a search, you paste or upload the brief and say:
Create this role in Kinetic from the attached brief, then show me who in my database matches.
The assistant reads the brief, creates the job with the title, requirements and responsibilities filled in, and runs the match, all in one go. Then "draft a shortlist summary I can send the client" gives you the write-up to review. One conversation, no tab switching, no re-keying the brief into two systems.
3. Chain your ATS with the tools you already have connected
This is the part that changes how the day feels. Your assistant does not just hold KineticRecruiter; it holds it alongside whatever else you have connected. If your email is connected in Claude, you can say:
Pull the must-have skills out of that thread with Acme and find candidates who match.
The assistant reads the requirements from the email and searches your database against them. Take notes from a client call and say "turn these into a job record in Kinetic", and the role is created. Ask it to "draft an update for Acme on where each candidate is right now", and it reads your live pipeline and writes the email in your inbox for you to check and send.
Notice what is doing the work here. We are not claiming KineticRecruiter integrates with your email or your calendar. Your assistant is the integration layer. It can hold your ATS and your other tools in the same conversation, and you decide how they connect. That is a far more flexible arrangement than a fixed list of native integrations, and it works with whatever you have connected today.
4. Ask in your own words, not "prompt engineering"
You are busy. You are not going to craft a careful prompt. You are going to type the way you would message a colleague:
who've we got thats done payments BA stuff in syd, contract ok
That works. The assistant reads the intent, searches your database, and comes back with matches. You do not need to learn a query language or care about AI at all. You need to find the person, and you asked for them the way you think about them.
Where it stops
Be clear-eyed about what this is. The assistant asks before it writes anything to your ATS: creating a role, moving a candidate to the next stage, updating notes, you see and approve the action first. It can only see and do what your Kinetic role allows, so a read-only user stays read-only through the assistant too, and every action lands in the audit trail. And it is a working tool, not an autonomous recruiter. It surfaces, drafts and files at your instruction. The judgment, and the send button, stay with you.
What the connector actually supports
So there is no guesswork, here is what the assistant can do through the connector today:
- Search and read: find candidates, jobs and clients; open full records; pull ranked match results with scores and reasons for a role; read pipelines, organisation stats and the audit log.
- Create, with your approval: add a candidate, create a job, add a candidate to a job or a pipeline.
- Update, with your approval: move a candidate through pipeline stages, update candidate notes.
Composing and sending emails, and anything outside those actions, is your assistant's own doing with your other connected tools, not something the connector reaches out and does on its own.
How to set it up
There is no install and no IT project. You connect your KineticRecruiter workspace to your assistant with one link, the same way you would add any other connector in Claude or ChatGPT. It works with both. If you are technical, Claude Code and Cursor connect too, but you do not need any of that to use it day to day. See the connector feature page for the walkthrough.
Which plans include it, and how that compares
The connector is included on every KineticRecruiter plan, at no extra cost, from Starter up. It is not a premium add-on and it is not gated to the top tier.
That is the honest difference worth knowing. Manatal, the main budget ATS, also ships an MCP server, but it is restricted to its Enterprise Plus tier at USD $55 per user per month and is currently in public beta (Manatal pricing, verified July 2026). As more ATS vendors add connectors, "who has one" will stop being a useful question. "Who makes you pay the top tier for it" will not. We put it on every plan because searching your own database from the assistant you already use should not be a luxury feature.
If plain-language search is the part that interests you, it is the same engine described in our piece on semantic search versus keyword search, just reached through the assistant instead of the app.
FAQ
Do I need to be technical to use it?
No. If you can add a connector in Claude or ChatGPT, you can use it. The technical options, Claude Code and Cursor, exist for developers, but the everyday use is plain-language chat.
Does the assistant change my data on its own?
No. It asks before it writes anything, creating a role, moving a candidate, updating notes, and you approve each action. It also only sees and does what your Kinetic role permits, and every action is logged.
Does it work with ChatGPT as well as Claude?
Yes. The connector uses the open MCP standard, so any assistant that supports MCP can connect, including Claude and ChatGPT.
Which plans include the connector?
Every plan, from Starter up, at no extra cost.
Ready to try it on your own database? Start your KineticRecruiter trial. Free for 7 days. Cancel anytime.