Update startup script comments for Ollama CPU-only configuration

Force Ollama to CPU, re-enable -np 2 with checkpoint fix, and web search detection

Move warmup before Pipelines/WebUI and add Ollama auto-retry to startup script

Fix RAG relevance filtering, add startup sequencing, and -np 2 for parallel slots

Disable prompt cache for Qwen 3.5 hybrid arch and document perf findings

Add --poll 0 to eliminate idle CPU spin, RAG timing logs, and warmup script

Tune RAG retrieval quality and add source citations

Document RAG validation results and retrieval tuning roadmap

Add RAG setup and Pipelines connection instructions to README

Implement Phase 2 RAG: ingestion pipeline, hybrid search, and Open WebUI integration

updated openwebui

Update README with additional resources

updated readme

Update README for clarity on project purpose

Add remaining llama.cpp startup scripts to configs/