Quickstart

Get to know Langflow by building an OpenAI-powered chatbot application. After you've constructed a chatbot, add Retrieval Augmented Generation (RAG) to chat with your own data.

Prerequisites

• A running Langflow instance
• An OpenAI API key
• An Astra DB vector database with:
  • An Astra DB application token scoped to read and write to the database
  • A collection created in Astra or a new collection created in the Astra DB component

Build the basic prompting flow

tip

If you prefer a pre-built flow, click New Flow, and then select Basic Prompting.

Continue to Run the basic prompting flow.

1. From the Langflow dashboard, click New Flow, and then select Blank Flow. A blank workspace opens where you can build your flow.

The Basic Prompting flow will look like this when it's completed:

To build the Basic Prompting flow, follow these steps:

2. In the components sidebar, click Inputs, select the Chat Input component, and then drag it to the canvas. The Chat Input component accepts user input to the chat.
3. In the components sidebar, click Prompt, select the Prompt component, and then drag it to the canvas. The Prompt component combines the user input with a user-defined prompt.
4. In the components sidebar, click Outputs, select the Chat Output component, and then drag it to the canvas. The Chat Output component prints the flow's output to the chat.
5. In the components sidebar, click Models, select the OpenAI component, and then drag it to the canvas. The OpenAI model component sends the user input and prompt to the OpenAI API and receives a response.

You should now have a flow that looks like this:

With no connections between them, the components won't interact with each other. You want data to flow from Chat Input to Chat Output through the connections between the components. Each component accepts inputs on its left side, and sends outputs on its right side. Hover over the connection ports to see the data types that the component accepts. For more on component inputs and outputs, see Components overview.

6. To connect the Chat Input component to the OpenAI model component, click and drag a line from the blue Message port to the OpenAI model component's Input port.
7. To connect the Prompt component to the OpenAI model component, click and drag a line from the blue Prompt Message port to the OpenAI model component's System Message port.
8. To connect the OpenAI model component to the Chat Output, click and drag a line from the blue Text port to the Chat Output component's Text port.

Your finished basic prompting flow should look like this:

Run the Basic Prompting flow

Add your OpenAI API key to the OpenAI model component, and add a prompt to the Prompt component to instruct the model how to respond.

1. Add your credentials to the OpenAI component. The fastest way to complete these fields is with Langflow’s Global Variables.

   1. In the OpenAI component’s OpenAI API Key field, click the Globe button, and then click Add New Variable. Alternatively, click your user icon in the top right corner, and then click Settings, Global Variables, and then Add New.
   2. Name your variable. Paste your OpenAI API key (sk-…​) in the Value field.
   3. In the Apply To Fields field, select the OpenAI API Key field to apply this variable to all OpenAI Embeddings components.

2. To add a prompt to the Prompt component, click the Template field, and then enter your prompt. The prompt guides the bot's responses to input. If you're unsure, use Answer the user as if you were a GenAI expert, enthusiastic about helping them get started building something fresh.

3. Click Playground to start a chat session.

4. Enter a query, and then make sure the bot responds according to the prompt you set in the Prompt component.

You have successfully created a chatbot application using OpenAI in the Langflow Workspace.

Add vector RAG to your application

You created a chatbot application with Langflow, but let's try an experiment.

1. Ask the bot: Who won the Oscar in 2024 for best movie?
2. The bot's response is similar to this:

_10

I'm sorry, but I don't have information on events or awards that occurred after

_10

October 2023, including the Oscars in 2024.

_10

You may want to check the latest news or the official Oscars website

_10

for the most current information.

Well, that's unfortunate, but you can load more up-to-date data with Retrieval Augmented Generation, or RAG.

Vector RAG allows you to load your own data and chat with it, unlocking a wider range of possibilities for your chatbot application.

Add vector RAG with the Astra DB component

Build on the basic prompting flow and add vector RAG to your chatbot application with the Astra DB Vector Store component.

Add document ingestion to your basic prompting flow, with the Astra DB component as the vector store.

tip

If you don't want to create a blank flow, click New Flow, and then select Vector RAG for a pre-built flow.

Adding vector RAG to the basic prompting flow will look like this when completed:

To build the flow, follow these steps:

1. Disconnect the Chat Input component from the OpenAI component by double-clicking on the connecting line.
2. Click Vector Stores, select the Astra DB component, and then drag it to the canvas. The Astra DB vector store component connects to your Astra DB database.
3. Click Data, select the File component, and then drag it to the canvas. The File component loads files from your local machine.
4. Click Processing, select the Split Text component, and then drag it to the canvas. The Split Text component splits the loaded text into smaller chunks.
5. Click Processing, select the Parser component, and then drag it to the canvas. The Parser component converts the data from the Astra DB component into plain text.
6. Click Embeddings, select the OpenAI Embeddings component, and then drag it to the canvas. The OpenAI Embeddings component generates embeddings for the user's input, which are compared to the vector data in the database.
7. Modify the Prompt component to contain variables for both {user_question} and {context}. The {context} variable gives the bot additional context for answering {user_question} beyond what the LLM was trained on.

_10

Given the context

_10

{context}

_10

Answer the question

_10

{user_question}

Adding variables like {context} creates new input handles in the component.

8. Connect the new components into the existing flow, so your flow looks like this:

9. Configure the Astra DB component.

   1. In the Astra DB Application Token field, add your Astra DB application token. The component connects to your database and populates the menus with existing databases and collections.
   2. Select your Database. If you don't have a collection, select New database. Complete the Name, Cloud provider, and Region fields, and then click Create. Database creation takes a few minutes.
   3. Select your Collection. Collections are created in your Astra DB deployment for storing vector data.
   info

   If you select a collection embedded with NVIDIA through Astra's vectorize service, the Embedding Model port is removed, because you have already generated embeddings for this collection with the NVIDIA NV-Embed-QA model. The component fetches the data from the collection, and uses the same embeddings for queries.

10. If you don't have a collection, create a new one within the component.

    1. Select New collection.

    2. Complete the Name, Embedding generation method, Embedding model, and Dimensions fields, and then click Create.

       Your choice for the Embedding generation method and Embedding model depends on whether you want to use embeddings generated by a provider through Astra's vectorize service, or generated by a component in Langflow.

       • To use embeddings generated by a provider through Astra's vectorize service, select the model from the Embedding generation method dropdown menu, and then select the model from the Embedding model dropdown menu.
       • To use embeddings generated by a component in Langflow, select Bring your own for both the Embedding generation method and Embedding model fields. In this starter project, the option for the embeddings method and model is the OpenAI Embeddings component connected to the Astra DB component.
       • The Dimensions value must match the dimensions of your collection. This field is not required if you use embeddings generated through Astra's vectorize service. You can find this value in the Collection in your Astra DB deployment.

       For more information, see the DataStax Astra DB Serverless documentation.

If you used Langflow's Global Variables feature, the RAG application flow components are already configured with the necessary credentials.

Run the chatbot with retrieved context

1. In the File component, upload a text file from your local machine with data you want to ingest into the Astra DB component database. This example uploads an up-to-date CSV about Oscar winners.
2. Click Playground to start a chat session.
3. Ask the bot: Who won the Oscar in 2024 for best movie?
4. The bot's response should be similar to this:

_10

The Oscar for Best Picture in 2024 was awarded to "Oppenheimer,"

_10

produced by Emma Thomas, Charles Roven, and Christopher Nolan.

Adding an Astra DB vector store brought your chatbot all the way into 2024. You have successfully added RAG to your chatbot application using the Astra DB component.

Next steps

This example used movie data, but the RAG pattern can be used with any data you want to load and chat with.

Make the Astra DB database the brain that Agents use to make decisions.

Publish this flow as an API and call it from your external applications.

For more on the Astra DB component, see Astra DB vector store.