Phase 4/5: add demo video (Git LFS), finalize README + docs

Phase 4: fix demo reset — --profile stream down -v wipes the Spark checkpoint

Phase 4: make record_demo.md commands runnable on Windows/PowerShell

Phase 4: unify — query the streamed proof-tier index with the shared search stack

Phase 3: Spark Structured Streaming -> Cassandra + Qdrant (containerized)

Phase 2: Kafka (Redpanda) broker + corpus producer

Phase 1: demo tier — FastAPI + Streamlit over a baked index, Docker-deployable

Phase 0: shared search core + reproducible corpus