Sitemap

Every page on RAG Explained Better, grouped by topic. The site is organised into eleven clusters — from the fundamentals of what RAG is, through the pipeline stages, the infrastructure and architectures that run it, and how to evaluate, secure and ship it. Start at the home page, or jump to any cluster below.

Fundamentals

• RAG meaning and definition
  • RAG use cases

Pipeline

• RAG chunking
  • Agentic chunking
  • Code chunking
  • Structure-aware chunking
  • Fixed-size chunking
  • Hierarchical chunking
  • Late chunking and contextual retrieval
  • Recursive chunking
  • Semantic chunking
  • Chunk size and overlap
  • Table chunking
• RAG context assembly
  • Contextual compression
  • Deduplicating retrieved context
  • Context ordering
  • Context window budgeting
• RAG embeddings
  • Embedding dimensions
  • Domain adaptation of embeddings
  • Embedding drift
  • Embedding model selection
  • Multilingual embeddings
  • Multimodal embeddings
  • Sparse embeddings
• RAG generation
  • Citations and attribution
  • Grounding the answer
  • RAG prompt design
  • Refusal and abstention
  • Structured output from RAG
• RAG indexing
  • Metadata filtering
  • Index updates and deletes
• RAG ingestion
  • Document cleaning for RAG
  • Source connectors for RAG
  • Incremental indexing
  • Metadata extraction for RAG
  • Scanned documents and OCR for RAG
  • PDF parsing for RAG
  • Tables and structured data in RAG
• RAG pipeline architecture workflow
  • RAG accuracy improvement
  • RAG pipeline build tutorial
  • RAG cost
  • RAG data preparation
• RAG reranking
  • ColBERT and late interaction
  • Reranking cost and latency
  • Cross-encoder reranking
  • LLM reranking
  • Reranker comparison
• RAG retrieval
  • Adaptive retrieval
  • Auto-merging retrieval
  • Contextual retrieval
  • Hybrid search
  • HyDE
  • Maximal marginal relevance
  • Multi-query and fusion
  • Parent-document retrieval
  • Query decomposition
  • Query rewriting
  • RAG-Fusion
  • Recursive retrieval
  • Query routing
  • Self-query retrieval
  • Sentence-window retrieval
  • Similarity metrics
  • Step-back prompting
  • Choosing top-k

Infrastructure

• HNSW
• IVF and flat indexes
• Multi-tenant RAG
• Vector quantization
• RAG infrastructure
  • Embedding models
    • Cohere embeddings
    • BGE and open-source embeddings
    • OpenAI embeddings
    • Voyage embeddings
  • RAG frameworks
    • Canopy
    • Cognita
    • Dify
    • DSPy
    • EmbedChain
    • Flowise
    • Haystack
    • LangChain
    • LangGraph
    • LlamaIndex
    • R2R
    • RAGFlow
    • Semantic Kernel
    • txtai
    • Verba
  • Knowledge graphs for retrieval
  • Vector databases
    • Azure AI Search
    • Chroma
    • Deep Lake
    • Elasticsearch and OpenSearch
    • FAISS
    • LanceDB
    • Marqo
    • Milvus
    • MongoDB Atlas
    • pgvector
    • Pinecone
    • Qdrant
    • Redis
    • SingleStore
    • Supabase
    • Turbopuffer
    • Vertex Vector Search
    • Vespa
    • Weaviate
• BM25

Architectures

• RAG architectures
  • Adaptive-RAG
  • Advanced RAG
  • Agentic RAG
  • Agentic RAG patterns
  • Cache-augmented generation CAG
  • Corrective RAG
  • FLARE
  • GraphRAG
  • Microsoft GraphRAG
  • Hybrid RAG
  • IRCoT
  • LightRAG
  • Long RAG
  • Memory in RAG
  • Modular RAG
  • Multi-agent RAG
  • Multi-hop multihop RAG
  • Multimodal RAG
  • Naive RAG
  • RAPTOR
  • ReAct RAG
  • Self-RAG
  • Speculative RAG

Evaluation

• Chunking evaluation
• RAG evaluation
  • RAG benchmarks
  • Generation metrics
  • LLM-as-a-judge
  • RAG evaluation metrics
  • Online evaluation
  • Ragas
  • RAG regression testing
  • Retrieval metrics
  • Synthetic evaluation data
  • Building a RAG test set
  • RAG evaluation tools
  • RAG tracing and observability

Failure modes

• RAG failure modes and problems
  • Chunk boundary loss
  • Conflicting sources
  • RAG debugging
  • Retrieval quality degradation
  • Enterprise RAG failure
  • Hallucination despite context
  • RAG latency
  • Multi-tenant leakage
  • Lost in the middle
  • Missing the obvious document
  • Multi-hop questions
  • Stale index
  • Numbers and tables
  • Vocabulary mismatch
  • Wrong chunk retrieved

Security

• RAG security
  • Access control
  • Audit logging
  • Compliance
  • Data poisoning
  • Data privacy
  • Document permissions
  • GDPR
  • Guardrails
  • HIPAA
  • OWASP risks
  • PII redaction
  • Prompt injection

Production

• RAG in production
  • A/B testing
  • Caching
  • CI/CD
  • Cost optimization
  • Deployment
  • Load testing
  • Monitoring
  • Observability
  • Rate limiting
  • Scaling
  • Serverless
  • Streaming
  • Versioning

Use cases

• Agent assist
• Enterprise chatbot
• Codebases
• Compliance Q&A
• Contract analysis
• Customer support
• Cybersecurity
• Documentation Q&A
• E-commerce
• Education
• Enterprise search
• Finance
• Government
• Healthcare
• HR
• Insurance
• Knowledge base
• Legal
• Manufacturing
• Report generation
• Sales enablement
• Scientific research

Decisions

• RAG decisions
  • Alternatives to RAG
  • Build vs buy RAG
  • Choosing a RAG framework
  • RAG vs CAG vs KAG
  • RAG vs context engineering
  • RAG vs fine-tuning
  • RAG vs knowledge graph
  • RAG vs long context
  • RAG vs MCP
  • RAG vs memory
  • RAG vs prompt engineering
  • Which vector database for RAG
  • When not to use RAG

Reference

• About
• Author
• RAG glossary
• Methodology
• RAG research index