Adapters

Database Adapters

QRyptoRAG supports multiple vector database backends through adapter interfaces:

• Supabase (pgvector): Production-ready with PostgreSQL and pgvector
• In-Memory: Perfect for development and testing
• File-based: JSON file storage for simple use cases
• Custom: Implement your own database adapter

Using Supabase

import { createClient } from '@supabase/supabase-js';
import { createSupabaseAdapter } from 'qr-video-rag';

const supabase = createClient(URL, KEY);
const database = createSupabaseAdapter(supabase);

Embedder Adapters

Generate semantic embeddings using your preferred provider:

• Google Gemini: Primary embedder, high quality and cost-effective
• Cohere: Multilingual embeddings with strong performance
• Hugging Face: Open-source models for privacy-focused deployments
• Google AI: Alternative Google embedding service
• Mock/Cached: For testing and performance optimization

Using Gemini Embeddings

import { createGeminiEmbedder } from 'qr-video-rag';

const embedder = createGeminiEmbedder(process.env.GEMINI_API_KEY);