NornicDB

NornicDB

Graph, vector, and historical truth in one database
Neo4j-compatible • Hybrid graph + vector retrieval • Historical reads via MVCC
Achieving Psygnosis for AI

Multi-arch support: CPU | CUDA | Metal | Vulkan

Quick Start • What It Is • Why NornicDB • Benchmarks • Features • Docs • Comparison • Contributors

Quick Start

Docker and Homebrew

# Homebrew
brew tap orneryd/nornicdb && brew install nornicdb && brew services start nornicdb

# arm64 / Apple Silicon
docker run -d --name nornicdb -p 7474:7474 -p 7687:7687 -v nornicdb-data:/data timothyswt/nornicdb-arm64-metal-bge:latest

# amd64 / CPU only
docker run -d --name nornicdb -p 7474:7474 -p 7687:7687 -v nornicdb-data:/data timothyswt/nornicdb-amd64-cpu-bge:latest

Open http://localhost:7474 for the admin UI. For an authenticated, loopback-only setup and your first query, follow the quick start guide. For NVIDIA CUDA hosts, use timothyswt/nornicdb-amd64-cuda-bge:latest. For Vulkan hosts, use timothyswt/nornicdb-amd64-vulkan-bge:latest.

---

Note: Docker on macOS does not expose Metal acceleration. The Apple Silicon image still runs, but GPU acceleration on macOS requires a native install from the releases page or a local build.

---

Local build

Install the build prerequisites, then clone, build, and start the server:

git clone https://github.com/orneryd/NornicDB.git
cd NornicDB
make build
./bin/nornicdb serve

On Windows, run .\bin\nornicdb.exe serve. See configuration for authentication and storage settings.

macOS release installer

Download the .pkg installer for your Mac from the release assets: choose arm64 for Apple Silicon or amd64 for Intel. Open the package and follow its installer prompts. Homebrew is also available above.

What NornicDB Is

NornicDB is a graph database that combines Cypher queries, graph traversal, vector search, and historical reads. It supports Bolt and HTTP interfaces and can be embedded in Go applications.

Why NornicDB Is Different

• Query connected data using Cypher and compatible Bolt drivers.
• Combine graph traversal with vector and full-text retrieval.
• Model historical state with MVCC and retained history.
• Run with CPU or supported GPU acceleration, or embed the database in a Go application.

See features, architecture, and the project details for capabilities, deployment patterns, and tradeoffs.

Performance Snapshot

Benchmark results depend on workload and hardware. Explore the performance guides, including Northwind, hybrid retrieval, and retrieval quality.

Features

• Cypher queries and functions and Neo4j migration.
• Vector search, memory policies, and agent integration.
• Transactions, historical reads, and durability.
• Embedding in Go, configuration, and deployment.

Documentation

Visit the documentation site for guides, API references, operations, and code exploration. The project details retain the full setup, features, benchmark, image, configuration, and roadmap information.

Comparison

Review the Neo4j compatibility guide, migration guides, and the documented comparisons when evaluating a migration.

Contributors

Contributions are welcome. See the contribution guide and contributors.

License

NornicDB is MIT licensed. See PATENTS.md for the defensive patent grant and NOTICES.md for third-party notices.

Psygnosis is a play on words or portmanteau meaning “mind" + "knowledge” in greek