Resume Parsing FAQ

We don't publish an accuracy figure until a benchmark backs one. Standard contact fields in a cleanly exported PDF are the easiest case; ambiguous layouts, infographic-style resumes and text embedded as images are the hardest. The most reliable way to judge it is on your own resumes: every account gets 50 free parses per month.

The API supports PDF files up to 5 MB. The parser extracts the text layer from the PDF, so it works with digitally created PDFs (exported from Word, Google Docs or LaTeX). There is no OCR, so scanned PDFs without a text layer cannot be parsed. DOCX and plain text are not supported; convert them to PDF before sending them to the API.

The response data contains: firstName, lastName, emailAddress, phoneNumber, location, locationCity, locationState, locationCountry, linkedInLink, workExperiences (title, organization, location, start and end month and year, isOngoing, and the description as markdown in skillsDescription), educations (degree, school, year, endYear, notes, location), fluentLanguages, and languages (each with a language and a level). Every field is optional, so check for missing values.

A parse can take up to 60 seconds, longer documents taking longer. The time includes PDF text extraction, AI field identification and building the JSON. The API is synchronous: you receive the full parsed result in the same HTTP response, so set your HTTP client timeout to at least 60 seconds. For user-facing flows, show a progress state while the request runs.

No, they are different steps in the hiring pipeline. Resume parsing extracts data from a document and structures it into fields like name, email, job titles and education. Resume screening takes that structured data and evaluates it against job requirements, scoring or ranking candidates. Parsing is a prerequisite for automated screening: you need structured data before you can apply matching logic. Our API does only the parsing step.

The parser reads resumes written in many languages, including English, Finnish, Swedish and other European languages. Field names in the JSON response are always in English regardless of the resume language, and month names are translated so dates come back as numbers. The languages and fluentLanguages fields capture the language skills the candidate lists. The PDF needs a text layer; scanned images are not supported.

Most of them. The parser handles multi-column designs, sidebars and tables because it reads text for meaning rather than position. Resumes where text is replaced by icons or embedded as images give weaker results, since there is no OCR. Standard professional formats give the best results.