LanceDB: Lightning-fast Vector Search & Seamless Scaling

LanceDB: Lightning-fast vector search & storage for AI apps. Build smarter tools with instant query responses, seamless scaling, and effortless integration—all without the hassle.

Visit Repository

✨ Research And Data

4.0(181 reviews)

271 saves

126 comments

38% of users reported increased productivity after just one week

About LanceDB

What is LanceDB: Lightning-fast Vector Search & Seamless Scaling?

LanceDB is a high-performance vector search engine designed for efficient similarity searches and scalable data management. Built for Node.js environments, it combines rapid vector operations with seamless scaling capabilities. By leveraging Ollama’s embedding models, LanceDB enables developers to quickly create and deploy applications requiring advanced semantic search and data analysis.

How to Use LanceDB: Lightning-fast Vector Search & Seamless Scaling?

Get started by cloning the repository and installing dependencies with `pnpm install`. To run searches, execute the provided test script via `pnpm test-vector-search` or directly through `node test-vector-search.js`. Configure your setup by pointing LanceDB to your storage path and connecting to Ollama’s API endpoint at `http://localhost:11434/api/embeddings`. For advanced use, integrate with MCP services by configuring the specified JSON snippet in your server settings.

LanceDB Features

Key Features of LanceDB: Lightning-fast Vector Search & Seamless Scaling?

Blazing-fast searches: Optimized algorithms ensure sub-second response times for large-scale vector databases.
Seamless scaling: Horizontally scale storage and processing without sacrificing performance.
Custom embeddings: Use Ollama’s pre-trained models or integrate your own via the `OllamaEmbeddingFunction` for 768-dimensional vector generation.
Efficient storage: Columnar data handling via Apache Arrow reduces memory overhead and speeds up queries.

Use Cases of LanceDB: Lightning-fast Vector Search & Seamless Scaling?

Power use cases such as:

Enterprise knowledge bases for instant document retrieval
Customer support systems with semantic query matching
Recommendation engines for personalized content delivery
Real-time analytics on streaming data with vector embeddings

LanceDB FAQ

FAQ from LanceDB: Lightning-fast Vector Search & Seamless Scaling?

What dependencies are required?: Node.js v14+, Ollama with `nomic-embed-text`, and write-accessible storage for LanceDB.
Can I use custom models with Ollama?: Yes, replace the default model in Ollama’s configuration to use alternative embeddings.
How does scaling work?: Leverage distributed storage systems like object stores or cluster setups for scaling beyond single-node limits.
What’s the performance like with large datasets?: Benchmarked at millions of vectors per second, with indexing optimizations for terabyte-scale data.

Content

LanceDB Node.js Vector Search

A Node.js implementation for vector search using LanceDB and Ollama's embedding model.

Overview

This project demonstrates how to:

Connect to a LanceDB database
Create custom embedding functions using Ollama
Perform vector similarity search against stored documents
Process and display search results

Prerequisites

Node.js (v14 or later)
Ollama running locally with the nomic-embed-text model
LanceDB storage location with read/write permissions

Installation

Clone the repository
Install dependencies:

pnpm install

Dependencies

@lancedb/lancedb: LanceDB client for Node.js
apache-arrow: For handling columnar data
node-fetch: For making API calls to Ollama

Usage

Run the vector search test script:

pnpm test-vector-search

Or directly execute:

node test-vector-search.js

Configuration

The script connects to:

LanceDB at the configured path
Ollama API at http://localhost:11434/api/embeddings

MCP Configuration

To integrate with Claude Desktop as an MCP service, add the following to your MCP configuration JSON:

{
  "mcpServers": {
    "lanceDB": {
      "command": "node",
      "args": [
        "/path/to/lancedb-node/dist/index.js",
        "--db-path",
        "/path/to/your/lancedb/storage"
      ]
    }
  }
}

Replace the paths with your actual installation paths:

/path/to/lancedb-node/dist/index.js - Path to the compiled index.js file
/path/to/your/lancedb/storage - Path to your LanceDB storage directory

Custom Embedding Function

The project includes a custom OllamaEmbeddingFunction that:

Sends text to the Ollama API
Receives embeddings with 768 dimensions
Formats them for use with LanceDB

Vector Search Example

The example searches for "how to define success criteria" in the "ai-rag" table, displaying results with their similarity scores.

License

MIT License

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.