Skip to content
Resumeboard

Lever resume parsing, which is thinner than the internet says

Lever's help centre has one short article on resume parsing. It is worth reading precisely because of how little it claims, and because what it does claim is specific and checkable.

The article that exists

Lever's Understanding Resume Parsing is a single paragraph. If you were expecting a specification of Lever's ranking or matching behaviour, it is not there, and it is not hidden elsewhere either.

The whole of it, quoted

"Resume parsing helps streamline and simplify the collection of candidate information. Lever's resume parser will extract information from the resume and add it to the candidate's profile, including their name, organization, and contact details. For full details on managing candidate files, refer to our help article on adding and deleting resumes."

Lever Support, Understanding Resume Parsing

Three fields are named: name, organization, contact details. That is the entire documented extraction surface. Note the asymmetry: it is "organization", not "employer and title and dates", and it is "contact details" as a group rather than separate email and phone fields.

The rules Lever states

Stated Lever resume handling rules
Rule What Lever says
Accepted file types "Our resume uploader supports most common file types for resumes, such as .doc, .docx, .pdf, and .txt." The word "such as" is doing real work here: this is an illustrative list, not an exhaustive one.
Maximum file size "The maximum resume file size you can upload is 10MB." Four times Greenhouse's 2.5 MB ceiling.
When parsing runs On candidate application, on manual resume upload from the candidate profile, and on referral submission. Lever's referral documentation notes that "Lever will parse the information from the resume and auto-populate the corresponding fields on the referral form."
Fields filled Name, organization, contact details. Nothing else is claimed.
Email parsing can be switched off "When the Skip email address parsing setting is enabled, the email address field in the contact information section of the profile will be empty for profiles created via candidate application on the Lever-hosted job site or manual resume upload. Candidate email addresses will need to be manually input instead."
Custom careers sites get no parsing Lever documents that a job site built with its Postings API "does not support auto-parsing of resumes to pre-populate applications." If a company's careers page is a custom Lever integration, no parse happens.

Sources: Adding and deleting resumes, Configure Your Lever-hosted Job Site, Lever Career Site Options, Submitting referrals, all on the Lever help centre.

Two settings that change the outcome without telling you

Two out of three Lever sites may not parse you at all

The "skip email address parsing" setting and the Postings-API job site option are both employer-side switches, and both are invisible to the candidate. The practical consequence: on Lever, the thing that most reliably reaches the recruiter is the form you fill in, not the file you upload. Type your details into the Lever form rather than trusting the pre-fill. It costs a minute and it is the only part of the process you control.

What Lever does not document

  • No failure modes. There is no "why did my resume not parse" article. You cannot learn in advance which layouts Lever struggles with.
  • No size or type list for parse specifically. The 10 MB and file-type list is about the uploader. Lever does not say the parser handles all of them.
  • No extraction rules for work history beyond "organization". Titles, dates and bullet text are not claimed as parsed fields.
  • No keyword, ranking or matching documentation at all. If a Lever job posting says "AI ranked", that is a separate feature from resume parsing, and Lever does not document how it scores.
  • No language list.

Reading the silence honestly

There are two ways to read Lever's thin documentation, and the honest answer is that we cannot tell them apart from outside.

The generous reading

Lever's product has always been a small-team tool, so the help centre is short. The parser may well handle a two-column layout fine; nobody wrote it down.

The cautious reading

Lever documents name, organization and contact details because those are what it reliably extracts. Everything else is shown to a human, which is why Lever's applications are short.

Under either reading, the advice is the same, which is why it is safe to give: write plainly, fill in the form, and do not build a resume around a parser nobody has described.

The version of this that fits on a checklist

  1. Any of .docx, .pdf or .txt, under 10 MB. A text-based PDF is the sensible default because it is the format the other five vendors on this site handle most predictably.
  2. Put your name, your employer and your email where a parser looking for those three things will find them first: top of the document, plain text, no text box.
  3. Spell out employers. "Organization" is the field being filled, and a bare brand name is a worse match than a full legal name.
  4. Type your details into the application form instead of trusting the pre-fill, especially your email address.
  5. Assume everything past name, organization and contact is going to be read by a person, if at all. That argues for a layout a person can scan in ten seconds — the classic template or the minimal template, not a dense one.

Lever is also the system where the resume matters least to the machine and most to the reader, which is the opposite of the folk wisdom. If you are done with this page, read Lever compared with Greenhouse next; it is the sharpest contrast in the guide.

Check the file you are about to upload

Resumeboard confirms the type, the size, and that the name, employer and contact details are all plain text in the body of the document — the three things Lever says it extracts.

Opens the builder. The check is under Check, on the ATS tab.

Open the Lever check Lever vs Greenhouse