fix: install tesseract-ocr-eng in Dockerfile and add grayscale contrast preprocessing for image OCR

fix: non-blocking async timeout for vision analysis with fast Tesseract OCR fallback

fix: robust OCR/image ingestion parsing and add one-click Supabase document deletion to Search Playground

fix: ColBERT dense fallback and increase Nginx proxy timeouts to 300s

feat: implement Hybrid Intent Detection (Keywords + LLM) for CRAG

fix: Nginx proxy for /advanced and LLM JSON intent classification

feat: upgrade RAG conversational intent classification to multi-lingual LLM

Security Hardening: Fixed path traversal, BM25 RCE, IDOR, and removed backdoor

Refactor LLM architecture to SOLID principles, add Gemini Judge, and self-reflective RAGAS logic

feat: enterprise multi-page RAG platform with GraphRAG, CRAG, Guardrails, and VectorDB Inspector

feat: Add RAG Execution Transparency, Technique Compatibility Matrix, and RAGAS Quality Dashboard

security: Remove hardcoded GLM API key from source files (use environment variable GLM_4_7_API_KEY)

feat: Implement GLM-4.7-Flash RAGAS Judge and LLM fallback (Qwen -> GLM-4.7-Flash)

fix: rebuild BM25 index from Supabase on serverless cold start

Fix file upload UI refresh and add Metadata Filters UI

Initial commit: Multimodal RAG Pipeline V3.0 with Fallback Logic