Vector Databases

A quick tour of the popular vector databases, what each one is best for, and a comparison of their strengths and weaknesses.

1. The Big Three (Specialized Native Vector Databases)

Pinecone

  • Type: Managed/serverless vector DB
  • Why Popular: Easiest setup, no infra, strong ecosystem (OpenAI, LangChain).
  • Best For: Fast production use, enterprise teams wanting fully-managed solution.

Weaviate

  • Type: Open-source hybrid search engine
  • Why Popular: Strong hybrid search (keyword + vector), modular, container-friendly.
  • Best For: Teams wanting open-source flexibility and high-accuracy hybrid retrieval.

Milvus / Zilliz Cloud

  • Type: Industrial-grade vector DB
  • Why Popular: Massive scalability (billions of vectors), mature, enterprise-ready.
  • Best For: Heavy workloads, on-prem/private cloud setups, large enterprises.

2. Developer-Friendly / Local

ChromaDB

  • Type: Lightweight, local-first
  • Why Popular: Simple (pip install chromadb), great for prototyping.
  • Best For: Local dev, notebooks, small-medium production workloads.

3. High-Performance Challenger

Qdrant

  • Type: Rust-based native vector DB
  • Why Popular: High performance, efficient memory usage, strong filtering system.
  • Best For: High-performance apps, Rust/Go developers, cost-efficient workloads.

4. Integrated SQL Option

pgvector (PostgreSQL)

  • Type: Extension for Postgres
  • Why Popular: No need for new DB; vectors + relational data in one place.
  • Best For: 90% of business apps already using Postgres.

Comparison Table

Database Type License Strengths Weaknesses
Pinecone Managed Service Proprietary Easiest setup, serverless, reliable Expensive at scale, closed source
Weaviate Native DB Apache Best hybrid search, modular Slightly complex setup
Milvus Native DB Apache Massive scale, enterprise-ready Heavy infra requirements
Chroma Embeddable Apache Best DX, local-first Not built for huge distributed scale
Qdrant Native DB Apache Rust-fast, great filtering, efficient Smaller ecosystem
pgvector Postgres extension Open Source Simple stack, convenient Slower at large-scale workloads
Adesh Tamrakar
SOFTWARE ENGINEER · VAULT

Notes, insights and random discoveries from a working engineer's vault - written for future me, published for you.