Fix: correct HF Space URL in README

Fix: add loguru to hf_space/requirements.txt

Fix: add loguru to requirements.txt

Polish: BGE-large embeddings, contextual retrieval, 142 tests passing, lint clean

Pre-download ML models during Docker build to eliminate cold-start latency

Add eval dashboard, persistent chat, and multi-doc comparison

Add streaming responses, multi-turn chat, and Second Brain

Security, bug fixes, and UX improvements

Lower similarity threshold 0.35→0.20, improve example questions

Strip whitespace from API key — httpcore rejects keys with trailing newlines

Fix upload status: warn on 0 chunks, only route to user_upload if chunks > 0

Fix ingest_document call: collection_name not collection

Add Streamlit config to enable file uploads on HF Spaces

Add file upload to Streamlit demo — users can query their own documents

Add A-RAG — hierarchical retrieval interfaces (Feb 2026)

Add Speculative RAG — Google Research 2024 (~51% latency reduction)

Add TTRAG — test-time compute scaling for RAG (ICLR 2025)

Fix hf_space: path detection, updated requirements, deploy script

Fix HF Spaces: run Streamlit on port 7860

Add HF Space card metadata to root README

Skip quality gate when LLM API unavailable in CI

Fix Claude API call: pass system as plain string, not cache_control block

Add missing CI dependencies (bs4, anthropic, langchain, pypdf, rich)

Remove Ollama from CI, use Claude backend throughout

Switch CI to Claude backend with API key secret

Fix CI badge URL and add eval gate setup note

Update project description in README.md

Fix eval pipeline and add real benchmark numbers

Initial commit — production RAG system