How our resume feedback works

1. Keyword overlap

The live percentage compares your resume text with the job description. The calculation runs in your browser and does not itself spend AI credits.

1. Lowercase both texts and replace punctuation other than letters, numbers, +, #, periods, hyphens and slashes with spaces.
2. Split words on whitespace, discard words of two characters or fewer and remove a fixed stop-word list, including common words such as “the,” “and,” “experience” and “team.” Repeated words count once.
3. Count a job-description word as matched if it occurs in the filtered resume words or anywhere as a substring in the lowercase resume text.
4. Divide matched job-description words by all retained job-description words, multiply by 100 and round to the nearest whole number.

With no retained job-description words, the calculation returns zero; with a blank job description, the interface is hidden. The unmatched-word list displays up to eight words, longest first. Length does not establish importance.

This is a lexical check. It does not recognize synonyms, verify your skills or judge whether a mention is meaningful. Substring matching can count incidental matches: “SQL” can match “NoSQL.” Short terms can be excluded. A high percentage can still describe a weak or inaccurate application. Interface colors are display conventions rather than pass or reject thresholds.

2. AI resume assessment

The separate assessment submits your resume text and job description to the application’s configured AI provider. It costs 3 credits per successful run and returns an overall assessment score, category scores, AI-identified matches and gaps, and editing suggestions.

• Keywords and hard skills
• Action verbs and impact
• Quantified achievements
• Relevance to the role
• Formatting suggestions from the submitted text

These numbers are model-generated judgments. The implementation does not specify fixed category weights, and the overall score is not documented as an arithmetic average of category scores. Results can vary across runs, inputs, providers and models.

The assessment has not been calibrated against hiring outcomes or validated as an employer ATS score. It analyzes text, not a rendered PDF: it cannot confirm column order, visual layout, extracted text order or how a particular employer’s parser handles your exported document.

Read the explanation and suggestions rather than chasing a number. Do not add skills, employers, dates, qualifications or performance metrics you cannot support. Follow the employer’s file instructions and review your exported PDF before submitting it.