feat: implement summarization engine with Pydantic schemas and map-reduce processing pipeline

feat: implement note generation API with background processing, summarization, and topic classification logic.

feat: implement local model inference with Qwen2.5 and mDeBERTa, and add recommendation system service.

revert: rollback project files to stable chat backend (commit f1332f7)

feat: implement YouTube note generation pipeline with task tracking and automated summarization.

feat: add topic categorization logic and implement YouTube note generation API routes

feat: implement Pydantic configuration management, add end-to-end pipeline testing, and include documentation for the recommendation module.

feat: enforce strict 7-category system and scaffold recommendations API endpoint

Refactor: Move Gemini API logic to backend, clean up Flutter UI, and secure credentials

fix gemini config

server v2