Run graph and vector memory backends for agents with NornicDB
Use NornicDB when an agent workflow needs a local or self-hosted graph, vector, and temporal database for GraphRAG, memory decay, historical reads, and Neo4j/Qdrant-compatible retrieval.
npx skills add agentskillexchange/skills --skill run-graph-and-vector-memory-backends-for-agents-with-nornicdb
brew tap --trust orneryd/nornicdb && brew install nornicdb && brew services start nornicdb) or run Docker with the upstream CPU or Apple Silicon images, exposing ports 7474 and 7687 and mounting persistent storage.Use NornicDB when an operator needs one backend for agent memory, graph retrieval, vector search, and historical truth instead of wiring separate Neo4j, Qdrant, embedding, and policy layers. The operator can run NornicDB with Homebrew or Docker, connect through Neo4j-compatible Bolt/Cypher and Qdrant-compatible retrieval surfaces, use managed embeddings and temporal MVCC reads, and lean on the upstream `docs/skills/` material for agent-ready query and memory policies. Invoke this when building or testing an agent memory or GraphRAG stack that needs graph plus vector semantics. The scope boundary is operating a retrieval and memory backend for agents; it is not a generic database catalogue entry or SDK listing.
What this skill actually does
Inputs and prerequisites: NornicDB server, Docker or Homebrew, Neo4j/Cypher-compatible client or GraphRAG application code.
Setup notes: Install with Homebrew ( brew tap –trust orneryd/nornicdb && brew install nornicdb && brew services start nornicdb ) or run Docker with the upstream CPU or Apple Silicon images, exposing ports 7474 and 7687 and mounting persistent storage.
Source and verification boundary: use https://github.com/orneryd/NornicDB as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a Multi-Framework workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.