ai-study-lab

AI Study Lab is a local-first learning product that turns a topic or PDF into a concise AI-generated lesson and helps job candidates prepare for a specific role. Version 0.1.0 is a focused public MVP built with FastAPI, Groq as its primary AI provider, an optional Gemini fallback, and a framework-free frontend. The project explores two practical uses of structured AI output: active study and grounded interview preparation. It favors short, interactive learning activities and evidence-based coaching over long generated pages or opaque match scores. The application includes separate Home, Study, Interview Coach, and Progress workspaces. The Study feature generates a lesson for any topic at Basic, Intermediate, or Advanced level, provides a concise explanation and focused key points, exactly 5 generated flashcards, and exactly 4 multiple-choice quiz questions with immediate feedback. The Interview Coach compares a job description with pasted CV/resume text or one CV/resume PDF, optionally including one LinkedIn Profile PDF and one Cover Letter PDF, and provides a qualitative fit assessment without a fake numerical match score. The application uses FastAPI REST endpoints backed by provider-independent structured output and Pydantic validation, in-memory PDF text extraction with no document persistence, local multilingual embeddings and ephemeral Chroma vector search for Study PDF retrieval, and Groq primary generation with an optional transparent Gemini fallback. It includes deterministic, live, RAG, and LLM-as-a-judge evaluation with a combined HTML report, minimal terminal-only request/provider telemetry, and a responsive HTML, CSS, and Vanilla JavaScript interface. The project is released under the MIT License and is an early public MVP with session-only browser state and no database or saved user content.