fix: pre-translate non-English chunks to English before LLM inference

fix: enable sampling — temperature=0.8 for MCQ, 0.4 for evaluator

fix: stronger MCQ/evaluator prompts — no duplicates, no self-reference, English-only, section 2 always present

fix: truncate runaway question generation to the first question

fix: force English output in all LLM prompts regardless of PDF language

feat: add MCQ mode with 4 options, client-side evaluation and per-choice explanations

feat: Off-Brand badge - custom FastAPI+HTML frontend, zero Gradio chrome

fix: tighten question prompt to avoid compound multi-part questions

feat: switch to MiniCPM4-8B, add robustness, score tracker, language selector

feat: initial project structure with all core modules