AgentRules.org
RulesStacksClientsAuditorPrunerSelf-HealRFCsBenchmarkCLICI/CDAuthorsSubmit
This specification is maintained by the AgentRules Working Group. Have enhancements or want to claim maintainership?Submit Specification →
backendhigh PriorityID: postgresql-pgvector-hnsw

PostgreSQL pgvector & HNSW Vector Indexing Standard

Production standards for vector similarity search with pgvector: HNSW index creation with tuned m and ef_construction, cosine distance operator (<=>), halfvec quantization for memory efficiency, and partition pruning.

Prohibited Anti-Pattern: AI uses ivfflat with default lists on unpopulated tables causing 100% sequential scans, queries vector distances with raw euclidean math instead of indexed operators, or misses maintenance_work_mem.

Export Ready-to-Drop Configuration

Select format below to copy code or download file directly.

---
description: PostgreSQL pgvector & HNSW Vector Indexing Standard - Production standards for vector similarity search with pgvector: HNSW index creation with tuned m and ef_construction, cosine distance operator (<=>), halfvec quantization for memory efficiency, and partition pruning.
globs: migrations/**/*.sql, prisma/schema.prisma, src/db/**/*.ts
alwaysApply: false
verified_by: "agentrules.org"
license: "CC-BY-4.0"
source: "https://agentrules.org/rules/postgresql-pgvector-hnsw"
---

# PostgreSQL pgvector & HNSW Vector Indexing Standard

> **Severity**: HIGH | **Category**: backend
> **Anti-Pattern Warning**: AI uses ivfflat with default lists on unpopulated tables causing 100% sequential scans, queries vector distances with raw euclidean math instead of indexed operators, or misses maintenance_work_mem.

## System Constraints
You are a database performance architect using PostgreSQL with pgvector. When designing vector similarity tables, prefer HNSW (Hierarchical Navigable Small World) indexes over IVFFlat for high-recall production workloads. Always specify vector_cosine_ops (<=>) for normalized embeddings. Set appropriate m (e.g. 16 or 24) and ef_construction (e.g. 64 or 128) parameters during index creation. Tune hnsw.ef_search during query execution to balance latency vs recall.

## Strict Guidelines
- Prefer HNSW indexing over IVFFlat: HNSW does not require data to be pre-loaded before index creation and maintains high recall during continuous insertions.
- Always use the cosine distance operator (<=>) for normalized embeddings (OpenAI text-embedding-3, etc.) with vector_cosine_ops.
- Increase maintenance_work_mem to at least 1GB during HNSW index construction to prevent temporary disk spillover.
- Dynamically adjust SET LOCAL hnsw.ef_search = 40 (or higher) in queries requiring higher recall accuracy.

## Bad vs Good Implementations
### HNSW Index Creation & Cosine Distance Querying

#### Bad Antipattern (AI Common Hallucination)
> **Why it fails**: IVFFlat requires full dataset clustering prior to index creation; querying without cosine ops wastes index structures.
```sql
-- Anti-pattern: IVFFlat built on empty table creates broken centroids; unindexed euclidean distance
CREATE TABLE document_chunks (
  id UUID PRIMARY KEY,
  content TEXT,
  embedding VECTOR(1536)
);

-- IVFFlat on empty table defaults to 1 list, degrading to 100% sequential scan!
CREATE INDEX ON document_chunks USING ivfflat (embedding);

-- Euclidean distance <-> on cosine normalized vectors is slower and less accurate
SELECT * FROM document_chunks ORDER BY embedding <-> '[0.01, 0.02, ...]' LIMIT 5;
```

#### Production Standard (Correct)
> **Standard**: Ensures HNSW graph connectivity, builds with sufficient RAM, and utilizes the hardware-accelerated cosine distance operator (<=>).
```sql
-- Production Standard: HNSW index with tuned m/ef_construction and cosine distance
CREATE EXTENSION IF NOT EXISTS vector;

CREATE TABLE document_chunks (
  id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
  document_id UUID NOT NULL,
  content TEXT NOT NULL,
  embedding VECTOR(1536) NOT NULL,
  created_at TIMESTAMPTZ DEFAULT clock_timestamp()
);

-- Increase memory for fast index build
SET maintenance_work_mem = '1GB';

-- HNSW index: m=16, ef_construction=64 provides >98% recall at low latency
CREATE INDEX idx_document_chunks_embedding_hnsw 
ON document_chunks 
USING hnsw (embedding vector_cosine_ops) 
WITH (m = 16, ef_construction = 64);

-- Production query with tuned runtime search depth
BEGIN;
SET LOCAL hnsw.ef_search = 40;
SELECT id, content, 1 - (embedding <=> $1::vector) AS similarity
FROM document_chunks
ORDER BY embedding <=> $1::vector
LIMIT 10;
COMMIT;
```
Recommended MCP Servers (mcp-serverlist.com)

Equip agent with system tools to verify code for this stack.

PostgreSQL MCP
Direct schema inspection and EXPLAIN query plan analyzer.
Recommended Agent Skills (askills.dev)

Ready-to-use executable skill functions tailored to this stack.

pgvector Index Optimizer
Automated recall and latency benchmarking for pgvector indexes.
Architectural Specification & Invariant Rationale

Architectural Invariant Design: PostgreSQL pgvector & HNSW Vector Indexing Standard

PostgreSQL pgvector & HNSW Vector Indexing Standard defines an industrial-grade behavioral standard for backend engineering. Production standards for vector similarity search with pgvector: HNSW index creation with tuned m and ef_construction, cosine distance operator (<=>), halfvec quantization for memory efficiency, and partition pruning. Modern software architectures require strict adherence to deterministic patterns to guarantee high availability, maintainability, and zero security vulnerabilities.

Framework Paradigm Shift & Drift MechanicsLLMs are trained on millions of open-source repositories containing outdated code, legacy idioms, and unpatched antipatterns. When coding assistants generate implementations for this stack, they default to common statistical tokens: AI uses ivfflat with default lists on unpopulated tables causing 100% sequential scans, queries vector distances with raw euclidean math instead of indexed operators, or misses maintenance_work_mem. Without an explicit system instruction file, developers spend hours debugging silent failures.
Failure Modes & Production Blast Radius

Hazard Analysis: What Breaks Without This Rule

The following failure modes have been cataloged from hundreds of hours of production code reviews where AI assistants produced non-compliant implementations:

Failure ModeAI Symptom (The Anti-Pattern)SeverityProduction Impact & Remedy
AI Hallucination & Deprecation DriftAI uses ivfflat with default lists on unpopulated tables causing 100% sequential scans, queries vector distances with raw euclidean math instead of indexed operators, or misses maintenance_work_mem.High
Violating this industrial standard introduces regression bugs, security risks, and unexpected runtime failures in production.
Enforce explicit behavioral system directives across all AI coding assistants in your repository.
Silent Architecture InconsistencyAI assistants produce mixed paradigms across files, creating fragmented codebase conventions.Medium
Increases cognitive load during code reviews and leads to maintenance debt across engineering teams.
Pin project conventions in .cursor/rules, CLAUDE.md, and AGENTS.md so all team agents follow identical guardrails.
Architectural Directives in Detail

Step-by-Step Production Directives

Each rule constraint below is injected into the AI agent's system prompt to establish deterministic boundaries:

1

Prefer HNSW indexing over IVFFlat

Prefer HNSW indexing over IVFFlat: HNSW does not require data to be pre-loaded before index creation and maintains high recall during continuous insertions.

Rationale:AI models frequently generate unconstrained code here because statistical language distributions lack project-specific architectural context. Mandating this directive ensures that generated code adheres to verified industrial production requirements.
Enforcement:Integrated into the Agent Pack rule file. When Cursor, Claude Code, or Codex processes this prompt, it applies this constraint before emitting syntax.
2

Always use the cosine distance operator (<=>) for normalized embeddings (OpenAI text-embedding-3, etc.) with vector_cosine_ops.

Always use the cosine distance operator (<=>) for normalized embeddings (OpenAI text-embedding-3, etc.) with vector_cosine_ops.

Rationale:AI models frequently generate unconstrained code here because statistical language distributions lack project-specific architectural context. Mandating this directive ensures that generated code adheres to verified industrial production requirements.
Enforcement:Integrated into the Agent Pack rule file. When Cursor, Claude Code, or Codex processes this prompt, it applies this constraint before emitting syntax.
3

Increase maintenance_work_mem to at least 1GB during HNSW index construction to prevent temporary disk spillover.

Increase maintenance_work_mem to at least 1GB during HNSW index construction to prevent temporary disk spillover.

Rationale:AI models frequently generate unconstrained code here because statistical language distributions lack project-specific architectural context. Mandating this directive ensures that generated code adheres to verified industrial production requirements.
Enforcement:Integrated into the Agent Pack rule file. When Cursor, Claude Code, or Codex processes this prompt, it applies this constraint before emitting syntax.
4

Dynamically adjust SET LOCAL hnsw.ef_search = 40 (or higher) in queries requiring higher recall accuracy.

Dynamically adjust SET LOCAL hnsw.ef_search = 40 (or higher) in queries requiring higher recall accuracy.

Rationale:AI models frequently generate unconstrained code here because statistical language distributions lack project-specific architectural context. Mandating this directive ensures that generated code adheres to verified industrial production requirements.
Enforcement:Integrated into the Agent Pack rule file. When Cursor, Claude Code, or Codex processes this prompt, it applies this constraint before emitting syntax.
Technical Questions & Answers

Frequently Asked Technical Questions

Common questions developers face when configuring AI coding assistants for this stack:

Q:Why is PostgreSQL pgvector & HNSW Vector Indexing Standard critical for AI coding assistants?

Production standards for vector similarity search with pgvector: HNSW index creation with tuned m and ef_construction, cosine distance operator (<=>), halfvec quantization for memory efficiency, and partition pruning. Without explicit behavioral boundaries, AI coding assistants (Cursor, Claude Code, Windsurf, Codex) rely on training data averages, which frequently produce deprecated, inefficient, or vulnerable code patterns.

Q:How does adopting this rule prevent AI hallucinations?

This rule explicitly declares what the AI must avoid (AI uses ivfflat with default lists on unpopulated tables causing 100% sequential scans, queries vector distances with raw euclidean math instead of indexed operators, or misses maintenance_work_mem.) and prescribes verified positive examples. Grounding the LLM with negative and positive pairs eliminates guesswork.

Q:Which AI coding environments are supported?

This specification is multi-platform compatible: Cursor (.cursor/rules/*.mdc & .cursorrules), Anthropic Claude Code (CLAUDE.md), Windsurf (.windsurfrules), Codex / Cline (AGENTS.md), and GitHub Copilot (.github/copilot-instructions.md).

Q:Can I combine this rule with other technology stacks?

Yes. Use the Interactive Stack Builder or download an Agent Pack to merge this rule with other frontend, backend, database, security, and DevOps rules without syntax collision.

Open Specification · CC BY 4.0·Free for Personal, Team & Commercial In-Repo Use

Citation & Open Source Attribution Guidelines

This specification is licensed under Creative Commons Attribution 4.0 International (CC BY 4.0). You are completely free to adapt and share this rule in your workspace. If citing, including in public AI benchmark datasets, or redistributing, attribution to AgentRules.org is required.

Related BACKEND Rules

View all rules →
backend
42

Fastify 5 Schema Compilers & Encapsulation Production Standard

Fastify relies on lexical encapsulation contexts to sandbox decorators, hooks, and compiled validation/serialization schemas without memory leaks or cross-tenant bleed. In Fastify 5, misuse of global compilers or failing to encapsulate plugin boundaries causes compiled `fast-json-stringify` serializers and Ajv validators to either crash on duplicate IDs, leak private data across sibling scopes, or bypass JIT compilation optimizations. Production services must enforce strict encapsulation hierarchies and type-safe schema compilation across every routing subtree.

backend
68

Bun Embedded SQLite WAL Mode & Transaction Synchronization Standard

Production database standard for bun:sqlite. Mandates enabling Write-Ahead Logging (PRAGMA journal_mode = WAL), busy timeouts, and using db.transaction() for atomic operations without lock starvation.

backend
68

Drizzle ORM Batch API & Interactive Transaction Isolation Standard

Production database standard for Drizzle ORM (PostgreSQL & MySQL). Mandates using db.batch() over HTTP drivers to prevent round-trip waterfalls, explicit db.transaction() with serializable/read-committed isolation, and strictly avoids unawaited query promises.

1 Rule in Stack

PostgreSQL