LokiVector is an innovative embedded database combining the simplicity of a document store with the advanced capabilities of vector search. Designed for modern AI applications, it provides fast, reliable, and durable data storage with a focus on crash recovery.

Key Features

• Document Store: A high-performance in-memory document database that supports dynamic queries with a MongoDB-like syntax.
• Vector Search: Utilizes the HNSW algorithm to perform efficient similarity searches on embeddings, enabling applications to leverage AI-driven data retrieval.
• HTTP Server: Offers a RESTful API that incorporates secure authentication, API key management, and configurable rate limiting for enhanced security and usability.
• Crash Recovery: Built for durability, LokiVector guarantees complete data recovery after unexpected crashes, maintaining the integrity of documents and vector indexes.
• Command-Line Interface (CLI): Enables effortless server management and quick projects setup through command-line commands.

Installation

To get started with LokiVector, the core library can easily be installed via npm:

npm install @lokivector/core

Basic Usage

A simple example showcasing how to create a database and run basic operations:

const loki = require('@lokivector/core');
const db = new loki('example.db');

// Add a collection
const users = db.addCollection('users');

// Insert documents
users.insert({ name: 'John', age: 30 });
users.insert({ name: 'Jane', age: 25 });

// Query
const results = users.find({ age: { $gte: 25 } });
console.log(results);

Performance Metrics

• Query Speed: Queries execute in less than 1ms when indexed.
• Vector Search Efficiency: Searches complete in under 0.5ms using the HNSW algorithm.
• Robust Crash Recovery: Comprehensive end-to-end tests validate the integrity of data after crashes, confirming 100% recovery rates.

Documentation

Extensive documentation is available, including guides for vector search, API endpoints, and architecture, allowing developers to implement advanced features seamlessly:

• Full Documentation Index
• Vector Search Guide

Community Support

LokiVector operates under the MIT License, providing flexibility for developers to use and contribute to the project. Users are encouraged to join discussions and contribute via GitHub for enhancements and troubleshooting.

For additional information about the release, visit the GitHub repository. LokiVector is geared towards developers aiming to build next-gen AI solutions that require reliable data storage capabilities.