SmartNotes Lite
Offline, on-device AI student companion powered by small language models
Core Features Implemented
- Zero-Internet Local Inference Engine: Packages 4-bit quantized small language models (GGUF) to run fully offline on standard 8GB RAM laptops
- Automated Course Slide Quiz Generator: Ingests local PDF slides and generates multiple-choice quizzes with instant grading and reasoning
- Key Concept Summarizer: Extracts core definitions, formulas, and revision bullet points directly from uploaded lecture documents
- Local Vector Search (Offline RAG): Searches through course notes with sub-second semantic retrieval without internet or cloud APIs
System Architecture
- llama.cpp execution runtime with AVX2 instruction optimizations for CPU-bound inference
- In-memory vector embedding store converting local document chunks without cloud GPU costs
- FastAPI backend serving a lightweight single-page React frontend with streaming SSE token responses
- Memory-capped process management ensuring RAM usage stays strictly below 2.4 GB
Have a project or role?
Let’s TalkGet In Touch