Voy

Voy is a WASM vector similarity search engine written in Rust. It's supported in non-Node environments like browsers. You can use Voy as a vector store with LangChain.js.

Install Voy

tip

See this section for general instructions on installing integration packages.

npm

npm install @langchain/openai voy-search @langchain/community

Yarn

yarn add @langchain/openai voy-search @langchain/community

pnpm

pnpm add @langchain/openai voy-search @langchain/community

Usage

import { VoyVectorStore } from "@langchain/community/vectorstores/voy";
import { Voy as VoyClient } from "voy-search";
import { OpenAIEmbeddings } from "@langchain/openai";
import { Document } from "@langchain/core/documents";

// Create Voy client using the library.
const voyClient = new VoyClient();
// Create embeddings
const embeddings = new OpenAIEmbeddings();
// Create the Voy store.
const store = new VoyVectorStore(voyClient, embeddings);

// Add two documents with some metadata.
await store.addDocuments([
  new Document({
    pageContent: "How has life been treating you?",
    metadata: {
      foo: "Mike",
    },
  }),
  new Document({
    pageContent: "And I took it personally...",
    metadata: {
      foo: "Testing",
    },
  }),
]);

const model = new OpenAIEmbeddings();
const query = await model.embedQuery("And I took it personally");

// Perform a similarity search.
const resultsWithScore = await store.similaritySearchVectorWithScore(query, 1);

// Print the results.
console.log(JSON.stringify(resultsWithScore, null, 2));
/*
  [
    [
      {
        "pageContent": "And I took it personally...",
        "metadata": {
          "foo": "Testing"
        }
      },
      0
    ]
  ]
*/

API Reference:

• VoyVectorStore from @langchain/community/vectorstores/voy
• OpenAIEmbeddings from @langchain/openai
• Document from @langchain/core/documents

Related

• Vector store conceptual guide
• Vector store how-to guides