Vector Databases

Purpose-built and general-purpose databases for storing and querying vector embeddings. Essential infrastructure for RAG, semantic search, and recommendation systems.

Quick Comparison

Tool Pricing Open Source Self-Hosted Maturity
Chroma freemium Yes Yes growing
LanceDB freemium Yes Yes growing
Milvus open-source Yes Yes established
pgvector open-source Yes Yes established
Pinecone freemium No No established
Qdrant freemium Yes Yes growing
Turbopuffer usage-based No No growing
Upstash Vector usage-based No No growing
Vespa open-source Yes Yes established
Weaviate freemium Yes Yes established

Chroma

growing

The developer-first open-source vector database

Open Source freemium Free Tier
Rapid prototyping (3 lines of code to start) Local development and testing Python-native AI/LLM workflows

LanceDB

growing

Serverless embedded vector database built on Lance columnar format

Open Source freemium Free Tier
Embedded vector search (no server needed) Multimodal data (text, images, video) Cost-efficient storage at scale

Milvus

established

Cloud-native vector database for billion-scale enterprise AI

Open Source open-source Free Tier
Billion-scale vector search GPU-accelerated similarity search Maximum index type diversity (12+ options)

pgvector

established

Vector similarity search for PostgreSQL

Open Source open-source Free Tier
Adding vectors to existing PostgreSQL Unified SQL + vector queries with JOINs and CTEs Avoiding new infrastructure

Pinecone

established

Fully managed vector database built for scale

freemium Free Tier
Zero-ops vector search at scale Serverless vector workloads Teams that don't want to manage infrastructure

Qdrant

growing

High-performance open-source vector database written in Rust

Open Source freemium Free Tier
High-performance filtered vector search Single-binary simplicity (no external deps) Multi-tenant applications

Turbopuffer

growing

Object-storage-first vector database with 10–100x cost savings

usage-based
Multi-tenant vector workloads Cost-efficient large-scale search Teams migrating from Pinecone

Upstash Vector

growing

Serverless vector database with true pay-per-request pricing

usage-based Free Tier
Serverless and edge architectures Pay-per-request vector search Low-traffic applications with cost sensitivity

Vespa

established

Hybrid search and ML serving engine for billion-scale applications

Open Source open-source Free Tier
Hybrid search + custom ML ranking at scale Billion-vector production deployments Real-time ML model inference alongside search

Weaviate

established

Open-source vector database with best-in-class hybrid search

Open Source freemium Free Tier
Hybrid search (BM25 + vector with Relative Score Fusion) Built-in vectorization at import time Multi-tenant applications with storage tiering