RAG Strategies

Strategy	Redwood	Cedar	Cypress	...
Speed	1-2 sec	2-3 sec	4-5 sec
Vector Search	✅	✅	✅
Chunking	✅	✅	✅
Memory	✅	✅	✅
Privacy	✅	✅	✅
Memory-enhanced prompt		✅	✅
Context-aware Query reframing		✅	✅
Document Priority			✅
Noice Reduction			✅
Reranking			✅
Data Normalization			✅

• Simplest/fastest approach, straightforrward RAG

• Best for clear, well-formed questions where the user's query is already optimal for retrieval

• ⚡ Fastest (~1-2 seconds)

• Best for ambiguous, context-dependent, or follow-up questions

• Slightly slower (~2-3 seconds) due to additional LLM rewriting step

• Better handling of conversational queries and ambiguous phrasing

• Tier-Based Source Retrieval

• Automatic Reranking

• Higher Retrieval Volume

• Equal Tier Treatment