nornicdb-hermes-memory-cpu
timothyswt/nornicdb-cpu-bge:latest
https://hub.docker.com/u/timothyswt
bridge
false
https://github.com/orneryd/NornicDB/issues
https://github.com/orneryd/NornicDB
NornicDB is a Neo4j-compatible graph + vector + temporal database and the best long-term memory store for the Hermes agent. This CPU-only build bundles BGE-M3 embeddings (no GPU required) and gives the Hermes agent sub-millisecond HNSW vector search, native graph traversal, and hybrid (vector + BM25) retrieval in one engine. Through Bolt/Cypher and Qdrant-compatible gRPC, the Hermes agent can store, auto-link, and recall memories without a separate vector store. Managed embeddings, auto-relationships, MVCC temporal/as-of reads, and policy-based memory decay let the Hermes agent build a persistent, self-organizing knowledge graph that forgets noise and promotes durable facts — ideal Graph-RAG memory for the Hermes agent. Connection guide (Hermes / Python / Node.js): https://github.com/RapalS/UNRAID_DOCKER_TEMPLATES/blob/main/docs/guides/nornicdb-connection.md
## NornicDB (CPU)
Graph, vector, and temporal database — Bolt/Cypher compatible with Neo4j drivers. **No GPU required.**
### Image variants
Pick a **version/branch** when adding the container:
| Branch | When to use |
|--------|-------------|
| **cpu-bge** | Default — BGE-M3 bundled (~586 MB). Web UI on port 7474 |
| **cpu-headless** | API-only, no web UI (~500 MB). Set `NORNICDB_HEADLESS=true` |
### After install
1. WebUI: port **7474** (cpu-bge only)
2. Bolt: `bolt://YOUR_UNRAID_IP:7687`
3. Health: `http://YOUR_UNRAID_IP:7474/health`
4. **Connection guide (Hermes / Python / Node.js):** https://github.com/RapalS/UNRAID_DOCKER_TEMPLATES/blob/main/docs/guides/nornicdb-connection.md
**Embeddings:** Do **not** bind-mount an empty folder to `/app/models` — that hides the bundled `bge-m3` model.
### Heimdall AI assistant (optional)
Enable natural-language chat in the admin UI (Bifrost). Set `NORNICDB_HEIMDALL_ENABLED=true`, then pick a provider:
- **local** — bundled GGUF model (default on *-heimdall images)
- **ollama** — set `NORNICDB_HEIMDALL_PROVIDER=ollama` and `NORNICDB_HEIMDALL_API_URL` (e.g. `http://YOUR_OLLAMA_IP:11434`)
- **openai** — set `NORNICDB_HEIMDALL_PROVIDER=openai` and `NORNICDB_HEIMDALL_API_KEY`
Full guide: https://github.com/orneryd/NornicDB/blob/main/docs/user-guides/heimdall-ai-assistant.md
For NVIDIA or Vulkan GPU on AMD64, use **nornicdb-hermes-memory-gpu** instead. For Apple Silicon, use **nornicdb-hermes-memory-apple-silicon**.
Tools:Utilities Network:Management
http://[IP]:[PORT:7474]
https://github.com/RapalS/UNRAID_DOCKER_TEMPLATES/blob/main/docs/guides/nornicdb-connection.md
Docker Hub pull. Recommend 2 GB+ RAM. No GPU.
https://raw.githubusercontent.com/RapalS/UNRAID_DOCKER_TEMPLATES/main/templates/nornicdb-hermes-memory-cpu.xml
https://raw.githubusercontent.com/orneryd/NornicDB/main/macos/Assets/NornicDB.iconset/icon_1024x1024.png
--restart unless-stopped
cpu-bge
CPU + BGE-M3 embeddings (~586 MB). Recommended default.
timothyswt/nornicdb-cpu-bge:latest
--restart unless-stopped
cpu-headless
CPU API-only, no web UI (~500 MB).
timothyswt/nornicdb-amd64-cpu-headless:latest
--restart unless-stopped
7474
7687
6334
/mnt/user/appdata/nornicdb/data
/mnt/user/appdata/nornicdb/export
/data
7474
7687
true
false
false
local
qwen3-0.6b-instruct
1024
0.5
-1
local
false