ATS readability: the file itself
Upload a PDF or DOCX to receive an immediate Aven readability score, without an AI request. These five original-file checks total 100 points:
- Text availability — 40 points. For PDFs, the proportion of pages with recovered letters or numbers. Pages with fewer than 25 non-space characters receive a separate inspection note; short pages alone do not lose points. For DOCX, usable recovered body text.
- Character recovery — 20 points, reduced in proportion to lines with replacement or control characters. No credit when no usable text is available.
- Recognisable sections — 20 points for two or more distinct recognised English or Arabic section groups; 10 for one.
- Contact method — 10 points for a readable email or international-format phone number.
- Date context — 10 points for a four-digit year. This does not verify its accuracy or association with a particular role.
Word tables, text boxes, multiple columns and header/footer text receive inspection notes, not automatic penalties. PDF column geometry, clipped text and visual overlaps are not measured by this score. Other languages may contain headings our rubric does not recognise.
Recovering a scan with OCR does not change the original file’s score. Edit the recovered content, export a fresh CV, then use Check exported PDF in the editor’s review panel, or upload that file again. A result for an earlier version is explicitly identified.
This is a transparent Aven heuristic. It is not a vendor-certified ATS test, a job-fit score, a factual verification or a prediction of a hiring decision. Your age, nationality, photograph and repeated keywords do not increase it.
The five writing checks
The weights add to 100. The same confirmed text and rubric version produce the same result. A high score means the text passes these checks; it is not a claim that the CV is finished or that every sentence is well written.
Recognisable sections · 20 points
Two distinct recognised section headings earn all 20 points; one earns 10.
Experience in context · 20 points
A recognised experience, project, education, research or volunteering heading earns 10 points. A four-digit year earns the other 10.
Direct contribution · 25 points
25 points multiplied by the proportion of narrative passages that do not start with a listed vague responsibility or generic self-description.
Readable passages · 20 points
20 points multiplied by the proportion of narrative passages containing 55 words or fewer.
Distinct evidence · 15 points
15 points multiplied by the proportion of narrative passages that are not exact repeats after normalising case and punctuation.
What is assessed separately?
The AI editorial review considers positioning and the evidence in your CV and produces passage-specific findings. A supplied job description can inform those findings, but job alignment does not contribute to this general text score. The saved-role matcher has its own evidence-fit method.
The text score does not inspect the original PDF’s layout, certify an ATS result, verify your facts, compare you with other candidates or predict an interview. Check the exported document and the employer’s requirements before applying.
When we withhold a score
We do not score very short or incomplete text, predominantly non-Latin text under this English rubric, or text with signs of unreliable extraction. OCR or partial recovery needs your confirmation first. You can correct the recovered text and review it again. Arabic and mixed-language document support does not make this English score valid for those CVs.
Fair interpretation
No points are awarded for a photograph, age, nationality, marital status, invented numbers or repeated keywords. Education and project evidence can support an early-career CV. We do not require everyone to have a long employment history or quantified outcomes.
Calibration and limits
This is an Aven-defined heuristic, tested with synthetic careers and controlled changes. It is not an independently validated hiring assessment. Editorial recommendations can reveal important weaknesses even when the text score is high. Use it to understand the checks, not to chase a perfect number.
Current rubric: text-review-2026-10-v1. Changes to criteria or weighting require a new version.