HawkinsDB provides a neuroscience-inspired memory layer for LLM applications, aiming to offer more human-like information storage and recall than traditional vector databases. It targets AI developers and researchers seeking to build more sophisticated AI systems by leveraging concepts from Jeff Hawkins' Thousand Brains Theory, enabling precise, context-aware queries and unifying semantic, episodic, and procedural memory.

How It Works

HawkinsDB utilizes "Reference Frames" as smart containers for information, capturing an entity's definition, properties, relationships, and context. This allows for nuanced queries beyond simple similarity matching. Knowledge is stored across multiple "Cortical Columns," mirroring the brain's ability to process information from diverse perspectives, thereby creating a richer, multi-dimensional understanding of data. It supports semantic, episodic, and procedural memory types and offers SQLite or JSON storage backends.

Quick Start & Requirements

• Installation: pip install hawkinsdb or pip install hawkinsdb[all] for full features.
• Prerequisites: Python 3.10+, OpenAI API key.
• Example: The README provides a Python snippet demonstrating HawkinsDB initialization, adding an entity with properties and relationships, and querying via an LLMInterface.
• Documentation: Examples are available in the repository's examples directory.

Highlighted Details

• Inspired by Jeff Hawkins' Thousand Brains Theory.
• Unifies semantic, episodic, and procedural memory.
• Integrates with ConceptNet for knowledge enrichment.
• Offers precise, context-aware querying over fuzzy similarity.

Maintenance & Community

The project is under active development, with a roadmap including multi-modal processing, performance optimizations, extended LLM support, and advanced querying. Contributions are welcomed via pull requests.

Licensing & Compatibility

HawkinsDB is released under the MIT License, permitting commercial use and integration with closed-source applications.

Limitations & Caveats

The project is currently under active development, with features like enhanced multi-modal processing and large-scale deployment optimizations still in progress.