What do I use for semantic search if I want minimal ops?

If you want minimal ops for semantic search, the simplest choices are usually:

1) Managed vector database

Best if you want the least infrastructure work.

Good options:

• Pinecone — very low ops, purpose-built for vector search
• Weaviate Cloud — managed version of Weaviate, easier if you want hybrid search
• Qdrant Cloud — simpler, fast, good cost/perf
• Zilliz Cloud (Milvus managed) — good at scale

Why: they handle indexing, scaling, persistence, backups, and availability for you.

---

2) Postgres + pgvector

Best if you already use Postgres and your scale is moderate.

• Add the pgvector extension
• Store embeddings alongside your app data
• Query with vector similarity directly in SQL

Why: very low operational overhead if you already run Postgres.
Tradeoff: not as specialized or scalable as a dedicated vector DB, but often “good enough.”

---

3) Search engine with vector support

If you already have search infrastructure:

• Elasticsearch / OpenSearch
• Use vector + keyword hybrid search

Why: good if you need both lexical and semantic search.
Tradeoff: more ops than pgvector, but still manageable if already deployed.

---

My practical recommendation

• Small/medium app, minimal ops: Postgres + pgvector
• Need managed, scale, and easiest ops: Pinecone or Qdrant Cloud
• Need hybrid keyword + semantic search: Weaviate Cloud or OpenSearch

If you tell me:

1. your rough document count,
2. expected query volume,
3. whether you already use Postgres,

I can recommend the best fit more precisely.