Lever Resume Checker: Fix Odd Characters and Encoding Noise
Lever's parser is commonly reported to be sensitive to unusual characters. Fancy bullets, ligatures, icon fonts and decorative symbols can turn into question marks or stray boxes in the parsed fields, and a name or email with a broken character is a problem for the person trying to contact you.
The tool above repairs broken character mappings, folds decorative characters to plain, safe equivalents, keeps the contact block extractable, and rebuilds the résumé as a clean, searchable PDF, DOCX or Markdown file. The audit scores the result against five ATS pipelines, Lever included. It is a local simulation, not a connection to Lever.
This is the ATS Resume Sanitizer set up for one case. See everything it does →
ATS Resume Sanitizer
Applicant Tracking Systems read a résumé as one linear text stream, and they fail in predictable ways: columns interleave, ligatures and broken fonts turn into garbage, content in form fields or images is dropped, and headers get truncated. The ATS Resume Sanitizer rebuilds your résumé as a single-column, machine-readable document, entirely in your browser. Your file never leaves your device.
What it does
- Spatial linearization: an XY-cut engine finds column gutters and block gaps and flattens columns, sidebars and visual tables into one reading order.
- Glyph & CMap repair: expands ligatures (fi, fl, ffi…), neutralises Private-Use-Area characters from broken CMaps, and folds exotic bullets, dashes and quotes to plain equivalents.
- Artifact stripping: the output is rebuilt from text only, so images, skill bars, decorative vectors, form fields and annotations are gone by construction.
- Semantic restructuring: content is sorted into canonical sections (Contact, Summary, Experience, Skills, Education, Certifications…).
- Dual-target output: a searchable text PDF plus an unstyled DOCX and Markdown, ready for any parser.
- Headless ATS audit: a built-in mock parser scores machine readability and projects it onto five pipelines: Workday, Taleo, Greenhouse, Lever and iCIMS.
A note on PDF/A
The PDF is a clean, searchable text PDF with ISO core fonts and a classic cross-reference table, which covers everything an ATS relies on. It is deliberately not stamped as PDF/A-1b/2b: true conformance needs colour profiles, XMP metadata and a validating toolchain that a client-side tool cannot guarantee. You get a document that parses correctly, not a badge it cannot verify.