Free ATS checker · 14 published checks
Check your CV.
Upload your CV for 14 checks on readability, structure and content. See the extracted text and the fixes to work on first.
- Your file stays on your device
- No account, unlimited checks
- Every weight published
Example result · Maya Collins
14 checks · model v4.5
78/100
Check score
A few things to fix
- Machine readability25 / 29
- Structure23 / 25
- Content18 / 30
- Format and hygiene12 / 16
- 2 critical
- 3 to improve
- 9 passed
Parser readback
In the file
Product Systems LeadLayer Labs · 2022 - Present
What the parser read
Product Systems LeadSkills - read out of orderLayer Labs · 2022 - Present
A sidebar heading landed between the job title and the employer.
Upload your CV
Drop your CV here
No CV yet? Start with an ATS-friendly templateYour CV file is processed on this device. Privacy details
- The file is read inside this tab. It is never uploaded or stored.
- The extracted text stays in this tab and is gone when you close it.
- Anonymous analytics, only if you accepted them: the 0-100 score and which mode ran, never CV content.
- AI rewrite, only if you press it: the single line you confirm is sent to write a suggestion, and is not stored.
No file needed. Layout checks need a PDF, everything else works from text.
No CV yet - paste the text or upload a file
The role you are applying for
Already in your CV
Missing from your CV
Add the ones that are genuinely true for you - never keywords you can't back up.
You get to see what the parser saw
Applicant tracking systems work with the text they extract from your file, not with the page you designed - and a layout can change the order that text arrives in.
Appliora shows you the exact text it scored. Below is a two-column CV example (Maya Collins), read by the same code that reads your upload.
The score tells you where to look. The readback shows you why.
See what to fix first.
The report points to the exact line behind each finding, so you fix the cause instead of guessing from a score.
- Ordered by impact, not by category
- The exact line behind every finding
- Re-check the file after you change it
Then match it to the role.
Paste the job description to see which requirements your CV covers, with the line that covers each one - and which are still missing.
14 checks. Every weight published.
The same model runs on every CV, including ones built in Appliora - our own templates get no bonus. Open a group to see exactly where its points come from.
Machine readability 29 points
-
A real text layer 10 pts
A scanned or exported-as-image CV contains no text at all. An ATS extracts characters - it finds none, so a recruiter's search has nothing to match. What an ATS actually is →
-
Reading order 14 pts
A two-column CV reads perfectly to you and comes out of a parser interleaved - job titles spliced into the middle of your skills list. Nothing in the text itself reveals this. Which CV format survives parsing →
-
Accented characters 5 pts
A recruiter searching for Świątek is doing a string comparison. If the export mangled your surname you are not ranked low - you are absent from the results.
Structure 25 points
-
Contact details 8 pts
Email, phone, city and LinkedIn have to be extractable as text. Contacts that live only in a header or footer are dropped wholesale by several parsers.
-
Recognised sections 10 pts
“Where I have been” is a lovely heading and no parser knows it means experience. Sections are matched against the names ATS software looks for. How to structure a CV →
-
Consistent dates 7 pts
Five date formats in one document means a parser gets your career length wrong, and an unexplained gap surprises you in the interview rather than before it. Explaining employment gaps →
Content 30 points
-
Bullet quality 13 pts
“Responsible for managing the social media team” describes a job description. The bullets that survive are the ones with a number, a result and a verb that did something. Putting numbers on your achievements →
-
Concrete skills 8 pts
Recruiters often search by tool and technology names. “Communication” and “MS Office” are too common to narrow a search. Which skills to list →
-
Spelling 5 pts
Only unambiguous misspellings, never your surname or your city. A dictionary run over a real CV flags the candidate's name more often than a genuine mistake, so this one reports from a curated list and stays quiet otherwise.
-
How it reads 4 pts
Six bullets that all begin “Managed” read as one block, and a recruiter's eye stops on the first word of each line. Verb variety, tense consistency and passive voice - measured, not guessed.
Format and hygiene 16 points
-
Length against experience 4 pts
Both directions are faults. Two pages from a graduate reads as padding; one page from fifteen years reads as a career with things left out. One page or two →
-
Format traps 6 pts
Tables, text boxes, skill rating bars and decorative glyphs. A parser reads none of them, so whatever they contain is simply gone. The mistakes that cost interviews →
-
Personal details 3 pts
Photo, date of birth, marital status, nationality. Normal in Poland and Czechia, a liability in the UK and US - so we tell you it is there and let you decide. Whether to put a photo on your CV →
-
File hygiene 3 pts
“cv final v3 (1).pdf” is the first thing a recruiter sees, next to forty other attachments, before a single word is read.
29 + 25 + 30 + 16 = 100 points
Model history
v4.5 · CurrentFixed a false reading-order flag on hanging-indent headingsView all versions
A score is only worth something if you can tell whether it moved because your CV changed or because we did. Every scoring change is versioned, and this is the record.
-
v4.5
No weight moved. The two-column check stopped mistaking a CV whose section headings sit closer to the margin than their own paragraphs for a sidebar. Both shapes produce the same well-populated gap in where lines start, but a heading that only ever precedes its own content, top to bottom, is not a second column - nothing about it can be read out of order. We found this auditing our own template collection against our own checker: three of our templates were being told their reading order was critically scrambled when it was not. Genuine sidebars - a skills or contact rail that runs beside unrelated content further down the page - are unaffected and still capped the same way.
-
v4.4
No weight moved. A photo on the first page is now reported as context and costs nothing. The check's own copy says a photo is normal and often expected in Poland and Czechia; a model cannot say that and take points for it in the same breath. What is worth knowing - anonymised recruitment removes it, UK and US employers generally prefer none, and no parser reads it - is still said. Overlapping roles were already reported without cost and are now filed the same way. The report itself changed shape: file actions live in one place, context findings sit apart from fixes, and the builder offers an editable copy rather than pretending an uploaded file already lives there.
-
v4.3
No weight moved. The fused-token check now asks for evidence that the parser did it. It used to count any twelve-character word with a capital in the middle of it, which is also the shape of every CamelCase employer a CV names: measured across our own export matrix, "PricewaterhouseCoopers" was reported as fused on all twenty templates and took 10 to 25 points off reading order on files with nothing wrong with them. A seam is now charged when a letter runs into a year ("Management2024"), when one long token carries two or more seams ("SeniorProductManagerWarsaw"), or - for a single seam - when that same file has already produced one of the first two. Genuine fusion is untouched: the two-column reference case still reports both of its fused lines, at the same severity and the same cost. Mangled encoding is no longer billed twice, once as damaged characters and again as a collided column.
-
v4.2
No weight moved. The parser now repairs what an exporter writes and nobody can see: soft hyphens inside words, zero-width and non-breaking spaces, the single-glyph "fi" and "fl" that LaTeX and InDesign emit, and accents stored as a letter plus a separate mark. That last one had been telling immaculate Czech and Polish CVs their diacritics looked stripped. Measured on the same document five ways: up to 19 points of movement, now zero. A running footer is also recognised when only its page number changes, so an email sitting in "Page 1 of 2" is flagged rather than credited. A .docx is now read the way a PDF is: headers and footers from their own parts, so a phone number that lives in a Word header is flagged rather than reported missing; the hyphen and the line break Word stores as elements are no longer deleted, which had been turning "Sales-Manager" into one unmatchable token; and the document title is read, so a leftover "Resume Template" is caught on a Word file as it always was on a PDF.
-
v4.0
A fourteenth check, How it reads: repeated opening verbs, mixed tenses, passive voice. Expectations now calibrate to career stage - a first CV is no longer told to quantify a summer job, and fifteen years without a number costs more than it used to. Weights moved to make room: Content is now 30 of the 100, up from 22 two versions ago.
-
v3.5
Headings are matched on what they contain, not on an exact string - so "Education and training", the wording Europass prints, is finally recognised. Sidebar detection stopped inverting on layouts where the fused lines are the body text.
-
v3.3
A CV whose reading order is critically scrambled can no longer score "ready to send". Polish misspellings match by stem, because the mistake lives in the stem and the ending is grammar.
-
v3.0
Column detection rebuilt. The previous model split the page down the middle, so the commonest CV layout on the market - a narrow left sidebar - was invisible to it, and a thoroughly scrambled CV scored 100 out of 100 on reading order. Spelling, language levels, orphaned skills and date attachment arrived with it, and every check now has to explain any point it takes.
A signal, not a guarantee
A high score means your CV did well across Appliora's published checks. It is not a hiring prediction: how systems parse, search and filter varies by employer, and a person usually makes the call.
No tool can promise a pass
Every employer configures its system differently, and many screen by hand.
Treat a high score as a strong technical green light and a low one as a specific, ordered to-do list - not a verdict on your career.
The same model for everyone
The engine does not know, and is not told, which tool built the CV it is reading.
A CV made in Appliora gets no bonus - and because every weight above is published, you can add the points up yourself.
The same file, the same score
The model is versioned.
Check a CV today and again next month without editing it and the number will not have moved - and if the model changes, the version printed under your score changes with it.
Which systems this is about
Employers may use global platforms like these, or ones common in Poland and Czechia. How a system parses, searches and filters a CV depends on how each employer has configured it.
- Workday
- SAP SuccessFactors
- Greenhouse
- Lever
- Taleo
- eRecruiter
- Traffit
- Teamio
Not partners or customers. Appliora does not simulate any one of these systems - the checks look for parsing problems that are common across them.
If you need a CV first
Start with a layout built to parse cleanly.
No CV yet - or too many layout findings? Start from a simple structure, add your content, then run the same checker before you send it.
Parsing bands come from running each template's PDF through this checker. They say how cleanly the layout reads, not how good a CV it will make.
Frequently asked questions
What is an ATS and why does it matter?
An ATS (applicant tracking system) is the software many employers use to collect, store and search applications. Recruiters search and filter candidates inside it, often before reading any CV in full. If the parser misreads your file - a scanned image, an unusual layout, a missing section - your details can arrive incomplete and become harder to find, however well qualified you are.
Is my CV uploaded or stored anywhere?
No. Your file is parsed entirely in your browser - it is never uploaded to a server, never written to disk, and no copy survives once you close the tab. Two things can leave the page, and only because you chose them. First, a single anonymous number: the 0-100 score and which mode ran, sent only if you accepted optional analytics. Second, if you press “Draft a stronger version”, that one line - and we show you the exact text and ask you to confirm before it goes. It is not logged or stored, and declining costs you nothing: the checks already ran on your device. If you choose “Create editable copy”, the recovered text is saved as a new CV in your browser, and synced to your account only if you are signed in.
What does the score mean?
The 0-100 score is Appliora's published model: how well your CV follows practices that make a file easy to parse and search - readable text, clear sections, contact details, measurable results, appropriate length and a simple layout. It is scored over the checks that could run on what you gave it: a PDF runs all 14, pasted text skips the layout checks and says so beside the score. No ATS produces a universal score like this one, and it is not a prediction of how any specific employer's system will behave.
Why did my CV get a low score even though it looks good?
Design and ATS-friendliness are different things. A visually polished CV can still score low if it's a scanned image with no real text layer, uses a multi-column layout that scrambles text order, or is missing a clearly labeled section.
Can this guarantee my CV will pass every ATS?
No tool can guarantee that. Employers use different systems and configure them differently, and a person usually makes the final call. This checker flags the most common, well-documented parsing pitfalls so you can fix them before you apply.
What should I do after checking my CV?
Fix the flagged issues one by one - add missing sections, add concrete numbers to your experience, and simplify the layout if needed. You can rebuild your CV from scratch in the Appliora builder, which uses ATS-friendly templates by default.
What counts as a good ATS score?
On this checker, 85 and up means the file is technically clean and ready to send - unless a critical finding is open, such as no email address, in which case the report says so above the number. 70-84 means a few fixable things; below 70, software is likely to lose or misread part of your CV before a person sees it. The thresholds are deliberately strict - a score that flatters you is not worth having. Other tools use their own models, so their numbers do not translate to this one.
Can an ATS read a PDF?
Yes, as long as the PDF has a real text layer - which it does if you can select the text with your cursor. A scan, a photo, or a design tool export that flattens everything to an image contains no text at all, so a parser has nothing to read. More on how an ATS reads your file.
Does an ATS reject a CV automatically?
Usually not on its own, although some employers set up knockout questions or filters. More often the problem is quieter: the parser misreads or drops part of your CV, so it does not match the recruiter's search and is harder to find in the results.
Do I need a different CV for every job?
Not a different CV - a different emphasis. The structure stays; the skills and phrasing move to match the posting. The Job Match tab shows exactly which terms from an advert are missing from your CV so you can decide which ones are genuinely true of you.
PDF or DOCX - which is safer?
PDF, if it was exported properly, because it looks the same everywhere and modern parsers handle it well. DOCX is a fine second choice and this checker reads both. The format matters far less than whether the layout is a single column. More on CV formats compared.
Why do Polish and Czech characters break in a PDF, and how do I fix it?
It happens when the font used for export has no glyphs for ą, ę, ł, ř or ž, or when the text is written in one encoding and read back in another. The letters arrive as question marks, empty boxes, or pairs like Ä… and Å‚. It matters because a recruiter searching for your surname is doing an exact character match, so a mangled name makes you unfindable. Fix it by exporting with Export as PDF rather than Print to PDF, using a font with full Central European coverage - Calibri, Arial, Times New Roman and Inter all work - and then opening the PDF and searching it for your own surname. If your own search cannot find it, a recruiter's search will not either. This checker tests for all three failure modes.
See what your CV looks like before you send it.
Run all 14 published checks in your browser and see exactly what the checker read.