Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

pgvector Semantic Search

An opinionated playbook that teaches Claude how to build and tune vector similarity search, RAG, and semantic search on PostgreSQL with pgvector.

Dev & CodingIntermediate★ 1,835⑂ 107AI score 9/10Last updated: Sep 4, 2026

What it does

Gives Claude a concrete, benchmark-informed default for vector search in PostgreSQL.

  • Golden path: halfvec(N) columns, cosine distance (<=>), HNSW with m=16/ef_construction=64, hnsw.ef_search=100
  • Index selection: HNSW by default, IVFFlat only for write-heavy or memory-bound cases
  • Quantization: halfvec first; binary quantization to bit(N) with re-ranking for very large sets
  • Filtered search: iterative scan, B-tree prefiltering, partial HNSW indexes, or partitioning
  • Ops: bulk COPY then index, maintenance_work_mem tuning, VACUUM/REINDEX, EXPLAIN ANALYZE, build-progress queries
  • A symptom → cause → fix table for the usual pgvector failure modes

Who it's for

  • Backend/AI engineers putting a RAG or semantic search layer directly on Postgres
  • Teams already on pgvector struggling with low recall or p95 latency spikes
  • Anyone sizing memory for millions of embeddings

Examples

  1. "Create a table and index for 1536-dim OpenAI embeddings" → produces halfvec(1536) DDL, HNSW cosine index, and an explicitly cast query pattern.
  2. "My filtered query returns fewer than 10 rows" → diagnoses HNSW early termination and suggests hnsw.iterative_scan = relaxed_order, max_scan_tuples, or partial indexes/partitioning.
  3. "20M vectors on a 64GB box" → recommends binary quantization with an 80× oversampling re-rank query (800 candidates → top 10).

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/timescale/pg-aiguide/HEAD/skills/pgvector-semantic-search/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the skills/pgvector-semantic-search folder from the GitHub repo timescale/pg-aiguide into my ~/.claude/skills/pgvector-semantic-search/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/timescale/pg-aiguide.git && mkdir -p ~/.claude/skills && cp -r pg-aiguide/skills/pgvector-semantic-search ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal.
  2. Clone the repository: git clone https://github.com/timescale/pg-aiguide.git
  3. Create the skills directory: mkdir -p ~/.claude/skills
  4. Copy the skill in: cp -r pg-aiguide/skills/pgvector-semantic-search ~/.claude/skills/
  5. Restart Claude Code and ask something like "design a pgvector semantic search schema" to trigger it.
  6. Before applying anything, confirm your database runs PostgreSQL 15+ with pgvector 0.8.0+ (CREATE EXTENSION IF NOT EXISTS vector;).
View source on GitHub ↗License: Apache-2.0