Greenhouse MCP: connect Greenhouse to Claude and ChatGPT
Ask Claude or ChatGPT how many applications a role has in each stage, which interviews are booked this week or what the panel wrote about a candidate. The answer comes straight from Greenhouse, without clicking through every job.
We build the connector for your company on Greenhouse’s own Harvest API. It only reads, your admins decide which groups may use it, and every call is logged. The AI summarises and finds; the people on the hiring team make the decisions.
Built for your company · Works in Claude, ChatGPT and Copilot
What your team can do with Greenhouse in Claude and ChatGPT
The pipeline in one answer
Ready for the debrief
This week’s interviews
Hiring next to the headcount plan
Ask it like this
“How many applications does the Senior Data Engineer role have in each stage?”
“Which interviews are scheduled for the Chicago office this week?”
“Summarise the scorecards for Tyler Brooks on the Product Designer role.”
“Which applications have been in the phone screen stage for more than two weeks?”
“Where did the candidates for the Account Executive opening come from this quarter?”
“Which open jobs in Engineering have no hiring manager assigned?”
What we build for Greenhouse
Deployed OS has no ready Greenhouse connector yet. We build it for your company when you start, on Greenhouse’s own API, with the same rules as every other connector: each person’s own login, access decided by group, every call logged.
- Jobs and openings
- Open and closed jobs with their departments, offices, hiring team and openings.
- Candidates and applications
- Applications per job with their current stage, source and dates, and the candidate details your company allows.
- Interviews
- Scheduled interviews with the role, time and interviewers.
- Scorecards and notes
- What interviewers have written, so the AI can summarise it for the hiring team. Offers and compensation stay out unless you ask for them.
- Departments, offices and users
- Your departments, offices and the people in Greenhouse, so the AI can say whose job or interview it is.
Who sees what
- Read-only
- The connector reads. It does not move a candidate, reject anyone, submit a scorecard or send an email.
- Your company decides who sees candidates
- Greenhouse’s Harvest API connects with company credentials a Site Admin creates, not each person’s own Greenhouse login. Access is set per group in Deployed OS and by the endpoints the credentials allow, so you decide which groups may see candidates at all, for example only recruiting and HR.
- People make the decisions
- The AI summarises and finds what is already in Greenhouse. It does not rank or reject candidates, and offers and pay stay out unless you choose otherwise, then only for the groups you name.
- Every call logged
- Who asked, what was read and when. Deployed OS keeps the log, and it can be exported.
Set up the Greenhouse connector in three steps
Greenhouse’s own MCP server or Deployed OS?
Greenhouse has its own MCP server, in open beta for Site Admins on Greenhouse’s newer plans. A Site Admin turns it on in Dev Center and chooses read and write access per area, and each person signs in with their own Greenhouse account, so the AI stays within their Greenhouse permissions. It can also take actions such as adding notes, moving applications and submitting scorecards. If you are on one of those plans and recruiting is what your team wants in its AI tool, it is a good fit.
Deployed OS is for when Greenhouse is one system among several, or you are on an earlier plan. We can build tools for your hiring process, such as a debrief summary in your format that combines Greenhouse with your calendar and HR system. Admins decide which groups get each system, every call is logged in one place, and the same rules apply in Claude and ChatGPT, next to your company’s skills and guardrails.
Greenhouse MCP: common questions
Yes. Greenhouse has its own MCP server in open beta, for Site Admins on Greenhouse’s newer plans, and it works within each person’s Greenhouse permissions. Deployed OS adds Greenhouse to one company connection with your other systems, with access per group and one log.
No. It summarises and finds what is already in Greenhouse, such as where an application stands or what the panel has written. The people on the hiring team read the candidates and make every decision.
No. The connector we build reads jobs, applications, interviews, scorecards and notes. It does not move or reject candidates, submit scorecards or send emails.
No. A company’s own integration with the Harvest API uses credentials your Site Admin creates for the organisation, so the connector cannot apply each recruiter’s own Greenhouse permissions. Access is set in Deployed OS instead: which groups get Greenhouse, and which endpoints the credentials allow at all.
Deployed is an EU company and fully GDPR compliant, and you choose where your data is stored, down to the city. Greenhouse data is fetched when someone asks for it.
It depends on the size of your team. Get in touch and we will walk you through it.
See it on your own tools.
Tell us which systems your company runs on. We’ll show you what your team can ask, and what IT gets to control.