Your Resume in Greenhouse, According to Greenhouse

Greenhouse’s help center says “Candidate uploads can be up to 100 MB”. Another article in the same help center, written for recruiters, says “Greenhouse Recruiting can’t parse resumes larger than 2.5MB.” No Greenhouse page we read connects the two numbers.

Greenhouse is an applicant tracking system, the software an employer uses to take in job applications and review them. The employer decides how it’s set up, and many of the features below are optional, limited to a subscription tier or sold as an add-on. Everything here comes from Greenhouse’s own public pages (the help center, the privacy policy, the product page and the AI principles page) as we read them on October 9, 2026. Most of the help center is written for employers, so it documents what an employer can configure and leaves open what the employer you applied to did configure. Offermic is not affiliated with Greenhouse.

What the parser trips on

A candidate can upload a .doc, .docx, .pdf, .rtf or .txt file, and the FAQ Greenhouse wrote for candidates has a preference: “Upload a PDF for best results.”

The 2.5 MB figure is in an article titled “Unsuccessful resume parse”. It’s written for recruiters who add a candidate from a resume file, and it lists formatting the parser can misread:

• “A resume with spaces between the letters.”
• “Resumes that include graphics, photos, or word art”
• “Resumes that are uploaded as an image, rather than a document (such as a .docx or .pdf)”
• “Complex resumes with tables, headers, and footers”
• “Resumes with the name and contact information in the header, footer, or text box”
• “Resumes that have a columned layout”
• “Resumes without clear sections and differing formats throughout each section”

One further entry is less expected: “Resumes with incomplete job titles. For example, Sr. Account Exec instead of Senior Account Executive”.

Put the FAQ and the article together and you get a text-based PDF under 2.5 MB with one column, your name and contact details in the body of the page, and job titles written out in full. Three checks in the file check of our resume builder come from that article: letter spacing, side-by-side columns and the 2.5 MB file size. That check and the resume score on the same page are our own measures. They are not a score from Greenhouse or any other applicant tracking system, and no employer sees them.

How far the list reaches is less clear, because Greenhouse describes resume parsing in three places and gives a different account in each. When a recruiter adds you by emailing your resume in, a route Greenhouse calls Maildrop, the parser pulls out “the name, email address, phone number, mailing address, current title, and the current company of the candidate”. For a MyGreenhouse profile, the candidate FAQ says parsing “is designed to save time, but can make mistakes” and doesn’t say which sections of the profile it fills. Talent Matching, the optional AI feature covered below, extracts “skills, job titles, years of experience, start/end dates of employment, and company names from employment history”. Greenhouse says that feature parses resumes with fine-tuned language models, and doesn’t say whether the list of misread layouts applies to it. Nor does it publish which profile fields, if any, are filled from the resume when you apply through a job post yourself.

No Greenhouse page we read says a failed parse rejects an application. When parsing fails outright, the article on failed parsing says, the recruiter sees a message that “the resume has only been attached to the candidate” and types the details in. Talent Matching treats an unreadable file the same way: “the candidate is surfaced to the recruiter for manual review”. We found no error rate or test result for the parser on the pages we read.

The five automatic rejections we found

We found five mechanisms in Greenhouse’s documentation that reject an application without a person deciding on that application. None is documented as reading the resume, and each is set up by the employer’s own admins or recruiters.

Mechanism	What triggers it	Availability Greenhouse states
Auto-reject application rule	Your answer to a yes/no, single-select or multi-select question on the job post	Plus and Pro tiers
Application limit rule	More applications than the employer allows in a set number of days, or re-applying to a job that rejected you before a set number of days has passed	Pro tier
Spam rejection	A later application from an email address after a user rejected one application from it as spam	Core, Plus and Pro tiers
Security risk rejection	A later application after a user rejected you for fraud, irregular application activity or misrepresentation	Core, Plus and Pro tiers
Spam blocklist	An email address, email domain or IP address on the employer’s list	Part of the Real Talent add-on

The first row is the one to remember while you fill in a form. Greenhouse describes auto-reject rules this way: “based on an applicant’s answer to a question, they will automatically be rejected as a potential candidate”. Among its examples are answers an employer treats as non-negotiable (“not enough experience, underage, requires a visa your company cannot provide, etc”). Nothing in the article has the rule look at the resume, and we found nothing in the documentation that lets the resume override it, so treat every yes/no and multiple-choice question as possibly decisive and answer accurately.

Fraud Detection, listed for the Plus and Pro tiers, isn’t in the table. It builds a report from your phone number, email address, location and IP address, and its own FAQ asks whether the feature makes automated rejection decisions and answers no. Recruiters can still filter flagged candidates out of search results with a checkbox labeled Exclude potential fraud.

We found no figures for the share of applications that are auto-rejected, filtered out or rejected in bulk, and no public list of employers using auto-reject rules or application limits. So nothing here can be set beside the claim that software rejects 75% of resumes. We traced that number in a separate piece, back to the oldest sources we could find.

Talent Matching, where an employer has switched it on

Talent Matching is the Greenhouse feature that rates resumes against a job, and it is optional twice over. Greenhouse lists it for the Core, Plus and Pro tiers with the Real Talent add-on, and a Site Admin then has to turn the AI on, for the offices and departments the admin picks. Greenhouse publishes no list of employers that use it. A sign we did find is a disclaimer about the feature on the job post, which the employer also has to turn on.

The hiring team begins by writing a calibration: skills, experience, job titles and optionally industry, each with an importance. The feature then sorts applicants into five categories: Strong match, Good match, Partial match, Limited match, or Needs manual review. In the recruiter’s table, candidates are grouped by that category, which the help center calls a match score, and ordered by application date within each group. We couldn’t find the thresholds between categories, or how the importance ratings in a calibration turn into one.

Only the resume goes into the match. Greenhouse’s data processing FAQ says “The matching algorithm is only using data coming from the resume”, and “Currently, structured application form data is not included in this calculation.” A skill you mention only in an application answer is outside the calculation, so put the evidence in the resume as well as the form. The same FAQ says names and contact information are not used for the matching, to avoid bias. It also says the parsed fields are “sent for external processing to third-party services (OpenAI)”, that “Personal data is not used to train our or third party’s models”, and that parsed resumes are cached for 30 days.

Matching isn’t limited to exact keywords. That FAQ says skills and job titles are represented “so that we can run a semantic search on them”. The Talent Matching FAQ adds that several terms can map to one calibrated skill, so “a longer list of matched terms doesn’t always mean a higher match score”. That’s a reason not to pad the list.

The recruiter gets the resume with matched terms highlighted, a short summary justifying the result and the matched and missing skills, and can override the category by hand. Greenhouse adds: “The AI categorization is only used during the initial application review stage and is not visible later in the process to prevent biasing interviewers.” No Greenhouse page we read describes showing it to the candidate. The documented route is a data access request to the employer, which can export a candidate packet containing the match category and reasoning.

On who decides, Greenhouse makes the same point in its help center, on its product page and in its AI principles. The help center’s wording: “Humans are required to make the decision for each candidate. There is no mechanism that automatically takes action to advance or reject a candidate based on an AI rating.” The documentation also shows bulk actions on the same table: “Select multiple candidates to advance, reject, or leave feedback on them.” So the software doesn’t reject by category, and a person can reject several applicants at once from a list that is grouped by category. Greenhouse doesn’t publish how often that happens.

An employer can let candidates opt out, which is optional. Where the opt-out is on, a “Learn more” link in the AI policy at the bottom of the job application leads to it. An opted-out application is labeled Needs manual review and goes to the team’s non-AI review, and Greenhouse deletes any match result that already exists while the application is active. The opt-out covers one application at a time. With no link, the candidate FAQ says, “the employer has not enabled this feature for the job you’re applying to”, and Greenhouse’s guide for employers leaves opt-out requests to the employer’s own process.

One caveat belongs next to all of this: an employer can connect third-party screening products to its Greenhouse account, and Greenhouse’s statements about Talent Matching don’t cover them. One integration article on the help center, which describes a partner’s product and sends support questions to that partner, says the product “reviews resumes against your job requirements”, can write a numeric resume score into the candidate’s Greenhouse profile and can “move candidates between Greenhouse stages when configured”. Greenhouse does not publish which integrations an employer uses.

The recruiter’s screen has your file on it

Every job’s interview plan starts with a stage called Application Review, where the job’s recruiter screens each new applicant, then advances or rejects the application. The help article on that stage lists what the screen holds: your documents, which can be viewed in the app or downloaded, your answers to the employer’s custom application questions, and an overview of other jobs you’ve applied to at that employer (for users with permission on those jobs). It doesn’t list a panel of parsed resume fields. As the article describes the screen, the reviewer has your file and your form answers in front of them.

Recruiters can narrow the queue by keyword. Talent Filtering, listed for the Core, Plus and Pro tiers, searches “in the full text of resumes and internal notes”, and the match is literal: “To appear in your results, the keyword from your search must exactly match the keyword in the application.” A recruiter can add “Greenhouse’s suggested keywords generated from your public job post”, which is a documented reason to name your main skills in the job post’s own words. The article describes a filter on the recruiter’s view and no action on the applications the filter leaves out.

Later interviewers can be given the file too. For a user with Basic permissions who is scheduled to interview you, the help center lists “Attachments (resume and cover letter)” among the things they can access, except attachments the employer marked Admin Only or Private. So be ready for questions about what’s in it. Our guide to resume-based interview questions explains how interviewers build them, and the resume to interview questions tool drafts the ones your own resume is likely to raise.

If you apply again

Greenhouse can merge a new application into an existing profile. The feature, auto-merge, is turned on by a Site Admin and matches on email address or phone number, optionally with further criteria. Where it’s on, the help center spells out what happens. Apply again to a job you’re still active on, and “Any new information will be absorbed into the older profile and they will continue to be managed like they never applied a second time.” Apply to a different job after a rejection, and the new hiring team “will see that the candidate was previously rejected while in Application Review”.

A different email address may not hide an earlier application. A Duplicate tag, separate from auto-merge, marks a new profile that matches an existing one. By default it checks email, phone number and LinkedIn URL, and an employer can add name. On the Pro tier, the application limit rule in the table above can reject a repeat automatically. All of that argues for applying once per job.

If you applied with the wrong email address, Greenhouse can’t change it. The candidate FAQ says to contact the employer “or re-apply with the correct email if the posting is still open”.

What you hear back, and from whom

A confirmation email is optional for the employer: “some organizations opt not to send these confirmation emails”. A rejection email isn’t guaranteed either, since the rejection window has a button labeled “Reject and don’t send email”. Recruiters select a rejection reason inside Greenhouse, and none of the candidate-facing features we read about displays it.

Greenhouse itself says it is solely a processor of applications for its customers and “will not share the status of any candidate application”. The data in your application is the employer’s to manage as well: the employer is the data controller and Greenhouse processes the data on its behalf. In the candidate FAQ’s words, “Greenhouse does not control any data you share with companies via your job application.” To see, correct or delete it, you ask the employer, who also sets how long the data is kept. Greenhouse publishes no default retention period, and its own article says “Deleting a candidate’s personal data must be done manually” once the retention timer has run out.

MyGreenhouse, the candidate’s side

MyGreenhouse is Greenhouse’s candidate portal, at my.greenhouse.io. What it shows about an application depends on what the employer chose to share there, and employer participation is optional. If the employer shares application status, a rejected application turns inactive after the rejection email is sent, or “4 days after the date of the rejection, if no email is scheduled or sent”. Greenhouse adds: “Candidates are not notified about the change”. If the employer has published stages, you see a candidate-facing stage name such as In review, and a rejection shows as Application closed. Applications made through LinkedIn, Indeed or other third-party sites may not appear at all.

A MyGreenhouse profile also autofills Greenhouse-hosted applications through Quick Apply. It holds one resume (“MyGreenhouse does not support multiple resumes. Your most recently uploaded resume is used by default”), and the FAQ tells you to review parsed information “before saving your profile and before submitting all applications”. So check the parsed profile and the single stored resume before each application.

Then there’s Dream Job. “You get one Dream Job each month.” Marking an application puts it in a priority queue: the article on Application Review says recruiters “can start by reviewing candidates who are referrals, internal applicants, agency submissions, or Dream Job applicants”. The employer doesn’t have to switch anything on, and “Dream Job selections cannot be undone”, so keep it for the application you care most about.

An account also changes who holds your data. “Greenhouse is the data controller for all personal information a candidate adds to their MyGreenhouse account”, and you can edit or delete the profile yourself. Deleting it doesn’t delete what employers already received. A profile is under different terms from an application. The privacy policy, last updated October 1, 2026, says Greenhouse may use it to “develop and train artificial intelligence (AI) models” and may “Provide your information to employers seeking candidates with your skills and background”. It also offers a link named Do Not Share My Profile with Potential Employers.

A MyGreenhouse account is the only part of this page where Greenhouse is the data controller. Everything else depends on how an employer has set Greenhouse up, and on whether the employer uses Greenhouse at all. For an employer on another system, we went through the public documentation of Workday, Lever, iCIMS and Ashby as well.