How the Job Description Keyword Scanner Extracts and Ranks Keywords (Heuristics, Not an LLM or a Real ATS)
The Job Description Keyword Scanner is a rule-based extractor that runs entirely in the browser with no AI, LLM, server or network call, and it is not a real applicant-tracking system. It takes a single pasted job description — no resume needed — tokenizes it, drops stopwords and recruiting boilerplate, and scores 1–3 word n-grams by how often they appear. It then boosts a curated set of hard-skill / tool / methodology terms (which weigh most because they matter most for ATS matching), a curated soft-skill bank, title and seniority words, capitalised terms (likely proper tools or skills), and any term sitting near required-context words such as “required”, “must have”, “proficient in”, “years of experience” or “degree”. The result is a ranked list, highest importance first, where each keyword carries an occurrence count and a flag for whether it appeared in a required context.
- Rule-based keyword extraction — no AI, LLM, server or real ATS
- Scores 1–3 word n-grams by frequency, boosting hard / soft skill banks, titles, capitalised and required-context terms
- Input is one pasted job description — no resume required
- Each keyword carries a frequency count and a required-context flag
- Hard skills and tools weigh most, then qualifications, titles and soft skills
