skills.rest
gtrig
gtrigCOMMUNITY
Explore all skills by this author →
GitHub

embedding-vector-search

Generate 1024-dim embeddings and perform HNSW similarity searches with gobed.

Updated Apr 11, 2026
Search Tags:#embeddings#hnsw#similarity#embed#vector-search#gobed
One-click install
npx skills add https://github.com/gtrig/LaightDB --skill embedding-vector-search
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: embedding-vector-search
Source: https://github.com/gtrig/LaightDB/tree/main/.cursor/skills/embedding-vector-search
Command: npx skills add https://github.com/gtrig/LaightDB --skill embedding-vector-search

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Generate embeddings for text and enable fast vector-based similarity search using gobed and an HNSW index for scalable internal retrieval.

Core Features & Use Cases

  • Pure Go embeddings using gobed for 1024-dim vectors
  • HNSW-based nearest neighbor search for fast retrieval
  • SavedGraph-based persistence for long-running indexes
  • Use Case: index internal documents and retrieve top-k similar items to a query

Quick Start

Load the embedding model, build the vector index, and perform a similarity search for a query.

Frequently Asked Questions about embedding-vector-search

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate text embeddings and perform vector similarity search in Go?▼

To generate text embeddings and perform vector similarity search in Go, you can use gobed to produce 1024-dimensional vectors and an HNSW index for fast nearest neighbor retrieval.

What is the best way to persist a vector index for long-running retrieval workflows?▼

Persisting a vector index for long-running retrieval workflows is handled through SavedGraph-based persistence, which supports HNSW graph insertions, searches, and graph state saving.

Does this vector search approach require a specific embedding model dimension?▼

Yes, this vector search approach requires the gobed model, which specifically generates 1024-dimensional embeddings to build and query the HNSW graph.

Can I use HNSW for internal document indexing and semantic matching?▼

You can use HNSW for internal document indexing and semantic matching to retrieve top-k similar items to a query, making it suitable for fast retrieval of snippets or conversations.

How do I retrieve the top-k most similar documents after generating embeddings?▼

After generating embeddings with gobed, you retrieve the top-k similar documents by loading the model, building the HNSW vector index, and executing a similarity search query against it.

What are the limitations of using HNSW and gobed for vector search?▼

A limitation of using HNSW and gobed for vector search is that the index requires SavedGraph-based persistence to maintain state, as the HNSW graph must be saved to support long-running insertions and searches.

Related Skills

Explore More Software Engineering Skills→
zepdbzepdb
43

zeppelin

Index and search text with Zeppelin vector store and BM25 full-text indexing.

Official
Advanced
justrachjustrach

emdb-search

Search EmergentDB for similar vectors and rank documents by similarity.

Community
Intermediate
kivo360kivo360
1

moai-essentials-refactor

Transform domain knowledge into vector embeddings for semantic retrieval.

Community
Intermediate
ccfccf

vector-index-tuning

Tune HNSW parameters and quantization for large-scale vector search indexes.

Community
Advanced