Short answer
Nextrova pulls 8 to 20 requirements from the job posting, checks each one against the text of your CV, weights them by importance, and combines that measured coverage with an AI assessment of how well your experience fits. Because the measured part is tied to your text, the same CV gives consistent results.
1. Requirements come from the posting
The posting sets the test. Nextrova extracts 8 to 20 concrete requirements: skills, tools, methods, certifications, languages, degrees and years of experience. Each one keeps the importance the posting gives it, from must-have to nice-to-have.
2. Each requirement is looked for in your CV's text
Matching runs on the text extracted from your file, so a requirement counts as found only if it is really there. Three rules keep the match honest:
- Whole words. Java is not found because your CV says JavaScript.
- Variants and abbreviations. JavaScript and JS, Kubernetes and K8s count as the same skill.
- Translations. A requirement can be found in another language, such as English written as Anglais in a French CV.
When the AI cites a passage as evidence for a match, that passage has to exist in your CV. If it does not, the citation is rejected.
3. Coverage is weighted and combined with an AI fit assessment
Coverage is weighted: a missing must-have lowers the score more than a missing nice-to-have. That measured coverage is combined with an AI assessment of how strongly your experience shows each requirement, for example whether a skill appears in a recent role with a result, or only in a list.
The AI models are Google's Gemini first, with other leading models through OpenRouter, Mistral and Groq as backups. Your name and contact details are removed before any of them reads your CV.
Why the same CV gives the same result
Because the measured part is anchored to your text, re-running the same CV against the same posting gives consistent results. When you edit your CV and run it again, the change in the score reflects the change in your CV.
What the ATS parse checks measure
Separately from the match, Nextrova checks the text extracted from your PDF or Word file:
- Contact details: email, phone and LinkedIn URL can be read.
- Standard headings: Experience, Education and Skills, recognised in English, French and Arabic.
- Dates that can be turned into a timeline.
- Readable characters: icon fonts and broken encodings that turn into symbols.
- Layout: columns, tables and text boxes that scramble reading order.
- Length and bullets.
What the score is not
There is no single ATS score that employers use, and every company configures its system differently. The match score is a measure of how clearly your CV shows what one posting asks for. It is not a prediction of whether you will get an interview, and no tool can promise that. For how other checkers define their scores, see what is a good ATS score.
What happens to your CV
Your CV is used to produce your analysis and stored so you can return to your reports. It is never sold and never shared with recruiters or employers, and your name and contact details are removed before the AI reads it. Guest analyses are deleted after 30 days.