What is the easiest way to run a vector database myself?

The easiest way depends on what you mean by “easy”:

1) Easiest overall: use a managed service

If you want the least setup and maintenance, use a hosted vector DB like:

• Pinecone
• Weaviate Cloud
• Qdrant Cloud
• Supabase + pgvector for smaller/simple use cases

This avoids running infrastructure yourself.

2) Easiest to self-host locally: Qdrant in Docker

If you want to run it yourself, Qdrant is one of the simplest options.

Run with Docker

docker run -p 6333:6333 -p 6334:6334 \
  -v $(pwd)/qdrant_storage:/qdrant/storage \
  qdrant/qdrant

Then open:

• REST API: http://localhost:6333
• Dashboard: http://localhost:6333/dashboard

Why Qdrant is a good “easy” choice

• Single binary/container
• Simple API
• Good performance
• Easy to persist data with a volume
• Works well for most RAG / semantic search apps

3) Easiest if you already use Postgres: pgvector

If you already have PostgreSQL, adding pgvector may be the easiest path:

• One database instead of a separate vector DB
• Familiar SQL
• Great for moderate-scale workloads

Quick recommendation

• Prototype / local dev: Qdrant with Docker
• Already on Postgres: pgvector
• No ops, production-ready: managed Pinecone/Qdrant/Weaviate

If you want, I can give you:

1. a 5-minute Qdrant setup,
2. a Docker Compose file, or
3. a Postgres + pgvector setup.