Skip to content

AI Knowledge and Retrieval: RAG, OKF, GraphRAG and Agent Memory

Nubenetes V2 Elite Portal

You are browsing the AI-Curated V2 Elite Edition. Looking for the exhaustive list of references? Check out the V1 Historical Archive.

Architectural Context

Detailed reference for AI Knowledge and Retrieval: RAG, OKF, GraphRAG and Agent Memory in the context of AI.

Deep-Dive Topic Pages

Enterprise AI Platforms ยท AI Agents & MCP ยท Core AI & LLMs ยท MLOps & GPU Scheduling ยท Databases

Architectural Comparison Matrix: RAG vs OKF vs GraphRAG vs Vectorless vs CAG vs Agent Memory

The 2026 Layered Knowledge Stack

Production AI systems in 2026 no longer treat Vector RAG as a one-size-fits-all default. Instead, architects combine Open Knowledge Format (OKF) for high-trust canonical rules in Git, Cache-Augmented Generation (CAG) for bounded static context, Vectorless Reasoning RAG (PageIndex) for dense structured documents, Contextual Hybrid RAG for massive unstructured archives, GraphRAG / LightRAG for multi-hop synthesis, and Persistent Agent Memory (Mem0, Graphiti, Letta, Cognee) for evolving session state.

Paradigm Core Mechanism Storage Backend Best-Fit Workloads Primary Trade-Off
Open Knowledge Format (OKF) Deterministic navigation of Markdown + YAML frontmatter (type, provenance, trust, freshness) + relative links Git repositories / plain files High-trust canonical knowledge: schemas, KPI definitions, runbooks, architecture rules Requires human or AI-assisted wiki curation; not suited for millions of raw unstructured logs
Traditional / Hybrid Vector RAG Chunking + Dense Embeddings + BM25 Hybrid Search + Reranking (Contextual Retrieval) Vector DBs (Qdrant, Weaviate, Milvus, pgvector) Massive, fast-moving, unstructured corpora (support tickets, docs, knowledge bases) Susceptible to semantic drift, chunk boundary loss, and continuous re-indexing overhead
GraphRAG and LightRAG LLM entity/relation extraction + graph community summaries + dual-level traversal Graph DB (Neo4j) or hybrid graph-vector stores Multi-hop relational queries and global thematic synthesis across large corpora Higher upfront indexing compute and LLM extraction cost (mitigated by LightRAG)
Vectorless Tree RAG (PageIndex) Hierarchical Table-of-Contents tree built over documents + iterative LLM reasoning search Tree JSON / document store (no vector DB) Dense structured documents (10-K filings, legal contracts, technical specs) Higher per-query reasoning latency than sub-50ms ANN vector lookups
Cache-Augmented Generation (CAG) Preload full bounded corpus into LLM long-context window and reuse KV / Prompt Cache Model KV Cache / Prompt Cache Bounded, stable reference corpora fitting within context limits (<200Kโ€“1M tokens) Bounded by model context window and cache TTL/storage cost
Persistent Agent Memory Fact extraction, bi-temporal validity windows, and agent self-edited memory blocks Temporal graph / hybrid store (Mem0, Graphiti, Letta) Long-running personalized assistants and stateful multi-session enterprise agents Needs explicit conflict resolution, privacy governance, and retention policies

The 2026 Hybrid Knowledge Stack Decision Framework

Choosing the Right Layer for Each Knowledge Workload

  1. Layer 1 โ€” Canonical Brain (OKF in Git): Store authoritative schemas, business rules, compliance policies, and SRE runbooks as OKF bundles versioned in Git with explicit YAML trust, provenance, and freshness signals.
  2. Layer 2 โ€” Bounded Static Context (CAG / Prompt Caching): When a reference manual or codebase subset fits within the prompt cache, preload it directly with CAG for zero retrieval error.
  3. Layer 3 โ€” Long-Tail Unstructured Search (Contextual Hybrid RAG / PageIndex): Use Contextual Retrieval (dense + BM25 + reranker) for high-volume archives, or PageIndex for exact hierarchical precision over dense PDFs.
  4. Layer 4 โ€” Multi-Hop and Temporal State (GraphRAG + Agent Memory): Layer LightRAG / GraphRAG for relational synthesis and Graphiti / Mem0 / Letta for time-aware agent memory across sessions.
  5. Layer 5 โ€” Live Operational Truth (MCP Tool Retrieval): Never embed volatile operational state; fetch live ground truth on demand via Model Context Protocol (MCP) servers.

Graphrag And Hybrid Knowledge Graphs

  • (2026) GitHub: HKUDS/LightRAG โญ 40812 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [DE FACTO STANDARD] [EMERGING] [GRAPHRAG] โ€” Fast, token-efficient dual-level graph-enhanced RAG framework combining low-level entity retrieval with high-level thematic synthesis at a fraction of traditional GraphRAG indexing and query cost.
  • (2026) GitHub: microsoft/graphrag โญ 36718 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [DE FACTO STANDARD] [EMERGING] [GRAPHRAG] [LLM] [RAG] โ€” Microsoft's modular graph-based Retrieval-Augmented Generation system combining LLM entity extraction, Leiden hierarchical community detection, and global map-reduce summarization for complex multi-hop reasoning.
  • (2026) GitHub: topoteretes/cognee โญ 33458 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [DE FACTO STANDARD] [EMERGING] [GRAPHRAG] [MCP] โ€” Open-source graph-vector hybrid memory engine using an Extract-Cognify-Load (ECL) pipeline to unify codebases, documents, and databases into traversable knowledge graphs for AI agents and MCP clients.
  • (2026) GitHub: neo4j/neo4j-graphrag-python โญ 1336 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [EMERGING] [GRAPHRAG] โ€” Official Neo4j Python package for building production GraphRAG pipelines combining Cypher graph traversals, vector similarity indexes, and automated knowledge graph construction.

Modern RAG Contextual Retrieval And Deep Document Parsing

  • (2026) GitHub: infiniflow/ragflow โญ 94532 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [DE FACTO STANDARD] [EMERGING] [RAG] โ€” Leading open-source enterprise RAG engine built on deep document understanding (complex layout analysis, tables, charts, OCR), grounded citations, and agentic retrieval workflows.
  • (2026) GitHub: run-llama/llama_index โญ 52981 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [DE FACTO STANDARD] [EMERGING] [RAG] โ€” Flagship open-source data and orchestration framework for Agentic RAG, hybrid retrieval, structured extraction, and multi-step knowledge workflows over enterprise data sources.
  • (2026) GitHub: deepset-ai/haystack โญ 27073 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [DE FACTO STANDARD] [EMERGING] โ€” Production-hardened, modular Python framework for composing hybrid retrieval pipelines (BM25 + dense vectors + rerankers), document ingestion, and Kubernetes-ready search microservices.
  • (2026) Anthropic Research: Introducing Contextual Retrieval ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [EMERGING] [RAG] โ€” Seminal engineering blueprint introducing Contextual Embeddings and Contextual BM25 (accelerated by Prompt Caching), reducing top-20 retrieval failure rates by 49% alone and 67% when paired with reranking.

Open Knowledge Format OKF And LLM Wikis

  • (2026) GitHub: AnswerDotAI/llms-txt (/llms.txt Specification) โญ 2699 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [EMERGING] [LLM] โ€” Minimalist Markdown standard (/llms.txt) for exposing curated, LLM-friendly documentation indexes at web roots so coding and research agents can discover authoritative context without brittle HTML scraping.
  • (2026) Google Cloud: How the Open Knowledge Format Can Improve Data Sharing (OKF v0.1) ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [EMERGING] [LLM] [OKF] โ€” Launch announcement from Google Cloud Data Cloud introducing OKF to formalize the LLM-wiki pattern and solve enterprise context assembly without proprietary vector lock-in.
  • (2026) Google Cloud: Introducing OKF v0.3 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [EMERGING] [OKF] โ€” Release breakdown of OKF v0.3 introducing refined conformance rules, explicit trust tiers, self-describing bundles, and interoperability with agent skills.
  • (2026) Google Cloud: OKF v0.2 Adds Trust Signals ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [EMERGING] [OKF] โ€” Architectural deep dive into OKF v0.2's five opt-in frontmatter trust primitives โ€” provenance, trust, freshness, lifecycle, and attestation โ€” enabling accountable AI-authored knowledge bases at scale.
  • (2026) Open Knowledge Format (OKF) Official Portal ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [EMERGING] [OKF] โ€” Official portal for the Open Knowledge Format (OKF), the Apache 2.0 vendor-neutral specification turning Git-versioned Markdown files with YAML frontmatter and hyperlinks into deterministic, auditable AI agent knowledge bundles.
  • (2026) GitHub: open-knowledge-format/open-knowledge-format โญ 125 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [EMERGING] [OKF] โ€” Official vendor-neutral specification (SPEC.md), conformance rules, reference bundles, and tooling for OKF (migrated from GoogleCloudPlatform/open-knowledge-format).

Persistent Agent Memory Systems

  • (2026) GitHub: getzep/graphiti โญ 32996 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [DE FACTO STANDARD] [EMERGING] [GRAPHRAG] โ€” Bi-temporal knowledge graph engine by Zep built for dynamic agent memory, tracking when facts became true and when they were superseded so agents reason accurately over evolving enterprise state.
  • (2026) GitHub: letta-ai/letta (formerly MemGPT) โญ 25511 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [DE FACTO STANDARD] [EMERGING] [LLM] โ€” Stateful Agent OS enabling LLMs to explicitly manage their own tiered memory hierarchy (in-context core memory vs archival recall storage) across long-running autonomous tasks.
  • (2026) GitHub: mem0ai/mem0 โญ 69337 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [DE FACTO STANDARD] [EMERGING] โ€” Universal memory layer for AI applications and autonomous agents that automatically extracts, consolidates, and retrieves cross-session user, session, and organizational facts via hybrid vector-graph storage.

Vector Databases Hybrid Search And RAG Evaluation

  • (2026) GitHub: chroma-core/chroma โญ 29724 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [DE FACTO STANDARD] [EMERGING] โ€” Developer-friendly, open-source AI-native search and embedding database supporting local prototyping and distributed cloud-native deployments.
  • (2026) GitHub: milvus-io/milvus โญ 46717 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [DE FACTO STANDARD] [EMERGING] โ€” Distributed, cloud-native vector database architected for billion-scale similarity search with disaggregated compute/storage and official Kubernetes Operator deployment.
  • (2026) GitHub: qdrant/qdrant โญ 35661 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [DE FACTO STANDARD] [EMERGING] [RAG] โ€” High-performance Rust vector and hybrid search engine featuring binary/scalar quantization, sparse + dense fusion, payload filtering, and native Kubernetes horizontal sharding.
  • (2026) GitHub: pgvector/pgvector โญ 23775 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [DE FACTO STANDARD] [EMERGING] [RAG] โ€” Open-source vector similarity search extension (HNSW and IVFFlat indexes) for PostgreSQL, enabling ACID-compliant hybrid SQL + vector retrieval without operating a separate vector database.
  • (2026) GitHub: weaviate/weaviate โญ 16959 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [DE FACTO STANDARD] [EMERGING] [RAG] โ€” Cloud-native AI database combining HNSW vector search, BM25 keyword scoring, multi-tenancy isolation, and modular generative search pipelines on Kubernetes.
  • (2026) GitHub: explodinggradients/ragas ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [EMERGING] [RAG] โ€” De facto open-source evaluation framework for measuring RAG and agentic retrieval quality across faithfulness, answer relevancy, context precision, and context recall in CI/CD pipelines.

Vectorless Reasoning RAG And Cache-augmented Generation CAG

  • (2026) GitHub: VectifyAI/PageIndex (Vectorless Reasoning-Based RAG) โญ 40734 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [DE FACTO STANDARD] [EMERGING] [LLM] [RAG] โ€” Leading vectorless, reasoning-based RAG framework that replaces chunking and vector databases with a hierarchical Table-of-Contents semantic tree navigated via LLM tree search for high-accuracy financial, legal, and technical analysis.
  • (2026) GitHub: hhhuang/CAG (Cache-Augmented Generation) โญ 1615 ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [EMERGING] [LLM] [RAG] โ€” Reference implementation of Cache-Augmented Generation (CAG), preloading bounded document corpora into LLM long-context KV caches to eliminate runtime retrieval latency and ranking errors.
  • (2026) arXiv: Don't Do RAG โ€” When Cache-Augmented Generation is All You Need for Knowledge Tasks ๐ŸŒŸ๐ŸŒŸ๐ŸŒŸ [EMERGING] [RAG] โ€” Foundational research paper benchmarking Cache-Augmented Generation (CAG) against standard RAG across multi-document QA benchmarks, demonstrating superior speed and accuracy for bounded knowledge bases.

๐Ÿ’ก Explore Related: Enterprise AI Platforms | AI Agents MCP | AI

๐Ÿ”— See Also: AWS | Azure