Frequently Asked Questions
Q.Is my pasted text sent to a server?
A.No. All calculations run in the browser. The text is only cached in your browser's localStorage so you can resume later. Nothing leaves your device, so copyrighted or confidential passages stay private.
Q.How accurate are the CSAT, TOEIC, and Lexile mappings?
A.Based on public formulas and research. Expected errors: ±1 CSAT band, ±50 TOEIC RC, ±80L Lexile. Official CSAT scores depend on question traps and yearly cutoffs too — our estimate is limited to the difficulty of the text itself.
Q.Why do SMOG and Flesch-Kincaid disagree on the same passage?
A.They weigh different signals. SMOG focuses on polysyllabic density, Flesch-Kincaid on sentence length + average syllables, Dale-Chall on vocabulary rareness. We average the grade-based scores for the CSAT mapping — individual scores tell you which axis is driving difficulty.
Q.How large is the CEFR dictionary and what's the source?
A.About 3,000 lemmas spread across A1-C2, curated by us by cross-referencing the Cambridge English Profile, Oxford 3000/5000, and the Dale-Chall word list. Inflections (plurals, past tense, 3rd singular, comparatives) auto-resolve to the root. Words not in the dictionary show in gray — typically proper nouns, technical terms, or rare inflections.
Q.Can I upload PDF or DOCX files?
A.Not yet — the MVP accepts pasted text only. PDF/DOCX upload, per-page analysis, and multi-passage batch compare are on the roadmap. For now, select the PDF text layer and paste it in.
Q.Why the 5,000-character cap?
A.To keep browser calculation sub-100 ms while typing. 5,000 characters cover two to three long CSAT passages. Beyond that, readability averages start masking per-section variation — analyze paragraph-by-paragraph for long docs.
Q.How do you pick the "hardest sentences"?
A.Each sentence gets a 0-100 difficulty score blending word count, avg syllables, number of subordinating/relative markers (that, which, because, etc.), and embedded comma count. Top 3 by score are shown with a short reason like "28-word sentence · 3 subordinate clauses".
Q.Does it support Korean passages?
A.English only for now. Korean syllable counting, POS structure, and Eojeol boundaries are incompatible with Flesch-style formulas. A dedicated Korean readability tool is on the roadmap.
Q.Login or paid plan?
A.None. Everything is free and ad-supported (Google AdSense). Optional paid features may appear later (cloud history, higher-accuracy CSAT/TOEIC model, classroom batch upload) but the core scoring stays free.
Q.Can I cite results in my blog or class materials?
A.Yes. Scores and mappings derive from public formulas; cite "readlvl.bal.pe.kr" and you may quote or screenshot freely. Please don't redistribute the bundled CEFR dictionary or iframe/mirror the site.