Vector Database Setup

Vector databases store data as high-dimensional vectors (embeddings), enabling semantic search — finding content by meaning rather than exact keyword matches.

Common use cases:

• RAG (Retrieval-Augmented Generation): Feed relevant documents to an LLM as context.
• Semantic search: Find similar products, articles, or support tickets.
• Recommendation systems: Suggest content based on user preferences.
• Deduplication: Detect near-duplicate entries in large datasets.

Workflows on Neura Market commonly integrate with:

• Pinecone: Fully managed, serverless. Great for getting started quickly.
• Weaviate: Open-source, self-hostable. Rich filtering and hybrid search.
• Qdrant: Open-source, high-performance. Excellent Rust-based engine.
• ChromaDB: Lightweight, embedded. Perfect for prototyping and small datasets.
• pgvector: PostgreSQL extension. Use your existing Postgres (including Supabase) as a vector store.
• Milvus: Open-source, distributed. Built for large-scale production workloads.

1. Choose a provider based on your scale, budget, and self-hosting preference.
2. Create an account or instance — most cloud providers have free tiers.
3. Get your API key or connection string from the provider dashboard.
4. Add to your workflow: Set the connection details as environment variables.
5. Create a collection/index: Define your vector dimensions (typically 1536 for OpenAI embeddings, 1024 for Cohere).
6. Ingest your data: Use the embedding + upload workflow included with many Neura Market products.