All modules Knowledge & Content

KAG — Knowledge Augmented Generation

Turn documents into a queryable knowledge graph — with semantic retrieval, multi-agent enrichment and graph analytics.

A single n8n node ("KAG Unified") that fronts the Knowledge Augmented Generation platform — a gateway over 11 backend microservices exposing 152 operations across 10 resources through a resource → operation picker. One node spans document ingestion and web scraping, knowledge-graph construction on Neo4j, semantic and hierarchical retrieval, LLM enrichment, plus literature discovery, scientific verification and multimedia search.

KAG combines HiRAG hierarchical decomposition and retrieval with KARMA multi-agent enrichment, writing entities and relationships into a Neo4j knowledge graph you can query with semantic search, Cypher or natural-language Ask. Beyond the core pipeline it adds academic paper discovery (snowball, citations, trust scoring), scientific verification (CoVe, RAGAS, formal proofs) and multimedia search with transcripts. Multi-tenant by design, it authenticates with a license token from your subscription portal.

KAG Unified API (license token)
Example workflow

Upload a folder of PDFs → KARMA enriches entities and relationships into a Neo4j graph → ask a natural-language question → return a cited, grounded answer.

IngestionEnrichmentAsk
What's included

Capabilities & operations

Ingestion & web scraping

Bring documents and the web into the platform.

  • Upload documents (sync, batch) with job tracking
  • Scrape URLs, bulk URLs and full sitemaps
  • Detect and list supported formats
  • Extract document and scrape images, screenshots and page renders
  • List ingested documents

Search & retrieval

Find and retrieve across your knowledge graph.

  • Semantic search and find-similar-nodes
  • Hierarchical (HiRAG) retrieval
  • Cypher, graph and optimized graph queries
  • Natural-language Ask

Processing & enrichment

Enrich and analyze text with KARMA, HiRAG and NLP.

  • KARMA document enrichment
  • HiRAG document decomposition
  • Entity extraction, parsing and segmentation
  • Language detection, quality assessment, similarity and entity density

Knowledge graph

Build and analyze a Neo4j knowledge graph.

  • CRUD for graph nodes and edges
  • Centrality, shortest paths and community detection
  • Bulk import / export
  • Graph statistics

Pipelines & ML

Run end-to-end pipelines and vector primitives.

  • Submit a full ingest → enrich → graph pipeline
  • Monitor, resume, cancel and delete jobs
  • Generate embeddings and compute attention / similarity
  • List available ML models

Academic discovery

Find and triage scholarly literature.

  • Search and semantic search across sources
  • Snowball expansion with citations and references
  • Paper trust scoring (single and batch)
  • Discovery sessions with insights, then push-to-ingest

Scientific verification

Reason over and validate the graph.

  • Topology: subgraph extraction, contradictions, authority, communities
  • Verification: CoVe, RAGAS and integrity checks
  • Formal methods: translate, verify and build proofs; sandboxed execution
  • Statistics: meta-analysis, confidence intervals and trend detection

Multimedia discovery

Discover and curate audio and video.

  • Multimedia search across sources
  • Transcripts, related items and quality scoring
  • Collections: curate, import and download
  • Push media into KAG
Who it's for

Teams building document intelligence, RAG and research assistants on top of their own knowledge graph.

Interested in KAG?

Purchasing is coming soon. Register your interest and we'll let you know when KAG and the rest of the collection are available.