A high-performance PostgreSQL extension for storing embeddings and performing vector similarity searches in AI applications.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install pgvector
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install pgvector using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is pgvector?
The pgvector skill provides a robust bridge between relational data and high-dimensional vector search. By integrating directly with PostgreSQL, it allows developers to store, index, and query embeddings alongside traditional metadata. This makes it an essential tool for anyone using Openclaw Skills to build sophisticated AI systems that require semantic understanding and efficient data retrieval.
pgvector Use Cases
- Semantic Search to find documents based on conceptual meaning rather than exact keyword matches.
- Retrieval-Augmented Generation (RAG) to provide contextually relevant data to LLMs for more accurate responses.
- Recommendation Engines that suggest similar items or products based on user behavior embeddings.
- Anomaly Detection for identifying outliers in high-dimensional datasets using distance metrics.
- Multimedia Search for querying image or video embeddings stored directly in your database.
How pgvector Works
- Database initialization establishes a connection to the PostgreSQL instance on port 5433.
- Vector table creation defines columns with specific dimensions (e.g., 1536) to match your embedding model.
- Indexing via HNSW or IVFFlat algorithms optimizes the database for high-speed similarity lookups.
- Embedding ingestion stores vector representations of text, images, or data alongside standard SQL types.
- Similarity querying uses specialized operators like cosine distance to retrieve the nearest neighbors for any given input vector.
pgvector Setup
To begin using this skill with Openclaw Skills, connect to your PostgreSQL instance and verify the extension:
# Connect to the local instance
psql -h localhost -p 5433 -U damien -d postgres
Enable the extension and create your first vector-enabled table:
CREATE EXTENSION IF NOT EXISTS vector;
CREATE TABLE documents (
id BIGSERIAL PRIMARY KEY,
content TEXT,
embedding vector(1536)
);
pgvector Data Schema & Taxonomy
The skill utilizes a structured schema within PostgreSQL to manage embeddings and associated metadata efficiently:
| Column | Type | Description |
|---|---|---|
| id | BIGSERIAL | Unique identifier for the document |
| content | TEXT | The source text or data used to generate the embedding |
| embedding | vector(N) | The high-dimensional vector storage (supports up to 16,000 dimensions) |
| metadata | JSONB | Extensible field for storing source info, authors, or tags |
| created_at | TIMESTAMPTZ | Automatic timestamping for record tracking |
pgvector Advanced Features
- HNSW (Hierarchical Navigable Small World) indexing for the highest accuracy and performance in similarity searches.
- IVFFlat indexing support for faster build times on exceptionally large datasets.
- Hybrid Search capabilities that allow for combining vector-based semantic scores with traditional full-text search rankings.
- Multiple distance operators including Euclidean distance (<->), Cosine distance (<=>), and Inner Product (<#>).
- Batch processing support using Python integrations for high-volume data ingestion pipelines.
SKILL.md
Loading
Related Openclaw Skills
Skill Manager
A centralized management utility to audit, categorize, and deduplicate your Openclaw Skills ecosystem.
Memory System Skill
A robust long-term memory management system for OpenClaw agents utilizing structured PostgreSQL and vector-based semantic search.
UI Design Optimizer
An automated design system generator that creates production-ready UI specs and starter code from local datasets.
Task Finish Contract
A protocol for AI agents to enforce explicit state tracking, evidence-based completion, and the elimination of mid-task stalls.
Skill Manager
A specialized management utility designed to list, categorize, and identify functional overlaps within your library of Openclaw Skills.
Video STT Skill
A robust utility for extracting audio from video URLs and converting it into accurate text or subtitles using local models or cloud services.
Featured*
