Cards / needs / Vector databases and AI search

Vector databases and AI search

Self-hosted alternatives to Pinecone, Weaviate Cloud, Zilliz Cloud and Qdrant Cloud.

What people say

• a vector database for my RAG app
• semantic search over our documents
• store embeddings without paying per query
• memory for my AI agent

Why people self-host it

People leave Pinecone, Weaviate Cloud, Zilliz Cloud, Qdrant Cloud for it. Why, strongest first: cost, privacy, control.

Does it suit a machine of your own?

Good fit fits well when the app using it is elsewhere; when it runs on the same machine, it need not be published at all

Questions that decide it

• how many documents or vectors, roughly?
• what will call it, and from where?

How to publish it

Usually

not published when only an app on the same machine uses it; otherwise its API on api-key, and its admin dashboard, if any, on org

When the data is private or sensitive

not published, or api-key only; embeddings can reveal the text they came from, so treat them like the documents

The machine

memory grows with the collection; fast disk

For builders

a free, private vector store for a prototype RAG or agent app, moved to a bigger machine later without changing the code

Cards for it

No card yet. GitHub repositories tagged vector-database, most starred and recently updated (Weaviate, Qdrant, Milvus, Chroma), and Postgres with pgvector when the app already uses Postgres An agent with the edgible-cards skill can shortlist apps and build one, or you can ask for one.