perf(pipeline): optimize OCR trigger, fail-fast multimodal metadata layer, and align frontend status indicator

feat(image-intel-v2): Simple English answers, connected Q&A layout, whitespace cropping, and things to remember appendix

feat(image-intelligence): Upgrade to V2 (metadata-driven scoring, programmatic post-solving head placement, inline HTML rendering)

Made the output answer structured

feat(backend): implement multi-intent question decomposition and adaptive composition

fix: serve favicon.ico and favicon.png explicitly in FastAPI and track favicon.ico with LFS

design: add custom favicon to browser tab and track with git lfs

design: add project logo and enhance README visuals for portfolio-grade presentation

docs: Redesign README and add MIT license for portfolio polish

Enhanced the project

Debug try 2

Fixed document download bug on deployment

Updated Docker file

Changes in readme.md for deployment

Merge Hugging Face remote

Made Gemini-3.1-flash-lite as primary model

Ready for deployment on Hugging Face

initial commit

Updated architecture and API quota limit

Dual mode update

Restored features from V1

Added frontend v1

Initial MVP backend complete