Quickstart with RAGStack for TS

This quickstart demonstrates a basic RAG pattern using RAGStack TS and the vector-enabled Astra DB Serverless database to retrieve context and pass it to a language model for generation.


RAGStack TS includes all the standard libraries you need for the RAG pattern, including the vector database, embeddings pipeline, and retrieval.

  1. Create a new project using NPM or Yarn:

    • NPM

    • Yarn

    npm init
    yarn init
  2. Then add the RAGStack package via the CLI:

    • NPM

    • Yarn

    npx @datastax/ragstack-ai install --use-npm
    npx @datastax/ragstack-ai install --use-yarn
  3. Set the AstraDB vector credentials. If you don’t have a vector database, create one at https://astra.datastax.com/.

    export ASTRA_DB_API_ENDPOINT=https://xx.apps.astra.datastax.com

    The Astra DB Serverless application token is associated automatically with the Database Administrator permission. An auth token example: AstraCS:WSnyFUhRxsrg…​).

    Both the endpoint and the token are available in the Astra Portal.

  4. Create an OpenAI key at https://platform.openai.com/ and set it as an environment variable:

    export OPENAI_API_TOKEN=sk-xx

RAG workflow

With your environment set up, you’re ready to create a RAG workflow in Javascript. Create a new file, index.js, and copy the following code:

const { OpenAIEmbeddings, ChatOpenAI } = require("@langchain/openai")
const { AstraDBVectorStore } = require("@langchain/community/vectorstores/astradb")
const { ChatPromptTemplate } = require("@langchain/core/prompts")
const { RunnableSequence, RunnablePassthrough } = require("@langchain/core/runnables")
const { StringOutputParser } = require("@langchain/core/output_parsers")

async function main() {
    // create the embeddings object with the OpenAI API key
    const embeddings = new OpenAIEmbeddings()

    // AstraDB connection parameters
    const astra = {
        token: process.env.ASTRA_DB_APPLICATION_TOKEN,
        endpoint: process.env.ASTRA_DB_API_ENDPOINT,
        collection: "demo",
        collectionOptions: {
            vector: {
                dimension: 1536, /** 1536 for OpenAI embeddings */
                metric: "cosine",

    /** Index some text into the Astra Vector Store */

    const vectorStore = await AstraDBVectorStore.fromTexts(
            "RAGStack is a framework for building RAG applications",
            "RAGStack has first-class support for AstraDB and Cassandra",
        [{source: "documentation"}, {source: "documentation"}],
    /** Now prepare the retrieval  */
    const prompt = ChatPromptTemplate.fromMessages([
        ["system", "You're an helpful assistant. Help the user to understand what is RAGStack. Use only information provided in the CONTEXT.\nCONTEXT:\n{context}"],
        ["human", "{question}"],

    const docParser = (docs) => {
        const formatted = docs.map((doc, i) => {
            return `<doc id='${i}'>${doc.pageContent}</doc>`
        return formatted

    const chain = RunnableSequence.from([
            context: vectorStore.asRetriever().pipe(docParser),
            question: new RunnablePassthrough(),
        new ChatOpenAI({}),
        new StringOutputParser()
    /** Finally ask a question about RAGStack to the chatbot */
    const answer = await chain.invoke("What is RAGStack?")
    console.log("Answer:", answer)

After that, you can run the script with Node.js:

node index.js
>Connected to Astra DB collection
>Answer:  RAGStack is a framework for building RAG applications. It also has first-class support for AstraDB and Cassandra.

Upgrade RAGStack version

After you have installed the RAGStack package, you can upgrade it to the latest version using the re-running the cli command:

npx @datastax/ragstack-ai install

or you can upgrade to a specific version:

npx @datastax/ragstack-ai install x.y.z

What’s next?

Was this helpful?

Give Feedback

How can we improve the documentation?

© 2024 DataStax | Privacy policy | Terms of use

Apache, Apache Cassandra, Cassandra, Apache Tomcat, Tomcat, Apache Lucene, Apache Solr, Apache Hadoop, Hadoop, Apache Pulsar, Pulsar, Apache Spark, Spark, Apache TinkerPop, TinkerPop, Apache Kafka and Kafka are either registered trademarks or trademarks of the Apache Software Foundation or its subsidiaries in Canada, the United States and/or other countries. Kubernetes is the registered trademark of the Linux Foundation.

General Inquiries: +1 (650) 389-6000, info@datastax.com